Wire Arc Additive Manufacturing of Scalmalloy^® (Al-Mg-Sc-Zr): Thermal Management Effects on Direct Age-Hardening Response

Abstract

The thermal history of a part deposited via wire arc additive manufacturing (WAAM) and hence its as-built properties can vary significantly depending on the thermal management applied, especially for metallurgically complex materials. Thus, this work aimed to assess the feasibility of processing thin-walled Scalmalloy^® (Al-Mg-Sc-Zr) structures by WAAM while examining the effects of arc energy and heat dissipation on their response to direct age-hardening heat treatment (without solution annealing). As a complement, the geometry, porosity, and processing time of such parts were also analyzed. The walls were built via the cold metal transfer (CMT) deposition process with different arc energy levels in combination with near-immersion active cooling (NIAC) settings (as thermal management solution), as well as with natural cooling (NC), resulting overall in both low surface waviness and porosity levels. Based on hardness testing, the resultant Scalmalloy^® direct-aging response (relative increase in hardness after direct age-hardening from WAAM as-built state) depended more on the arc energy per unit length of deposit applied. In contrast, the other thermal management approaches (NIAC or NC) helped in maintaining Sc in a supersaturated solid solution during deposition. Thus, Scalmalloy^® strengthening was demonstrated as feasibly triggered by means of a post-WAAM direct age-hardening heat treatment solely. Additionally, in comparison with a thermally equivalent (same interpass temperature) condition based on NC, the NIAC technique allowed the achievement of such a positive result on direct-aging response with much shorter WAAM processing times, therefore improving productivity.

1. Introduction

Scalmalloy^® is a scandium (Sc)- and zirconium (Zr)-modified aluminum–magnesium (Al-Mg) alloy which achieves high mechanical strength with low anisotropy through Al₃(Sc,Zr) precipitation and very fine grain formation [1]. In the context of additive manufacturing (AM), although such an advanced light alloy has received more attention for laser powder bed fusion (L-PBF) operations [2], it can also be processed via directed energy deposition (DED) techniques [3,4]. But despite promising results within the specific field of DED processes, including with deposition technologies based on the use of the electric arc as fusion source, there is a gap in terms of specific information regarding the effects of applying more effective/productive thermal management solutions on the direct age-hardening (without solution annealing) response of Scalmalloy^® parts when processed by wire arc additive manufacturing (WAAM).

Thermal management can be defined as any solution implemented, individually or in combination (hybrid cases), to aid with the way the heat is delivered to and/or dissipated from the parts being built in AM and hence to control their processing temperature [5]. This way, the thermal history (including heat accumulation) can be controlled and resultant features such as geometry/aspect, macro/microstructure, and mechanical properties, as well as production time, might be regulated for quality and productivity in WAAM.

The thermal management options for WAAM are diverse. The type of implementation, means of application, location of application, timing of action, and thermal effect all directly influence the outcomes. So, these key aspects may be beneficial to certain quality factors of the parts being produced but can also be unfavorable to others [5]. Therefore, selecting the appropriate approach to deal with the heat in WAAM can be more complex than it initially appears.

Table 1. Types of thermal management (TM) for WAAM: concepts and examples of application.

TM	Natural	Intrinsic (process-based)	Passive	Active
Concepts [5]	Relies on convection, conduction, and radiation for heat dissipation, with cooling times between layers during fabrication	Adjusts deposition parameters (such as power, wire/powder feed rate, and travel speed) and/or deposition path to control heat input during fabrication	Acts over the substrate and/or building platform to modify heat dissipation from the parts via conduction during fabrication	Applies direct means to the parts, such as by liquids/gasses, for more efficient temperature control during fabrication
Examples	Silva et al. [6] controlled the interpass temperature by the cooling time between the layers during WAAM of parts made from an ER309LSi wire	Zheng at al. [7] controlled the heat input in PTA 1-DED parts made from Inconel 625 powder, improving geometrical accuracy and reducing deposition time and energy consumption	Lu at al. [8] used a water-cooled building platform to avoid heat accumulation and improve geometrical quality in WAAM from an ER70s-6 wire	Reisgen et al. [9] applied water bath and water aerosol to cool thin-walled structures produced by WAAM from an ER70s-6 wire

¹ Plasma transferred arc (PTA) deposition process, which uses powder as feedstock material.

summarizes the basic concepts and exemplifies applications of the different types of thermal management approaches for such a manufacturing route.

According to a previous work on thin-walled Al structures built by WAAM [10], regardless of the thermal management technique applied, the geometry of the parts being built tends to be very similar if the same interpass temperature (IT) is maintained. However, to achieve the same IT level, the thermal history (peak temperatures, cooling rate, exposure time at elevated temperatures, etc., evolved from each approach possibly applied to deal with the implicated heat) might differ significantly. Thus, the as-built microstructure that is formed may end up being quite distinct depending on the existing thermal management scenario.

Wang et al. [4] verified the effects of thermal management on an Al-Mg-Sc-Zr alloy processed via laser-directed energy deposition (L-DED) by using air cooling (AC condition—for natural thermal management) versus a substrate with water cooling (WC condition—for passive thermal management). After direct age-hardening, the yield strength of a sample from the part produced under the WC condition was significantly enhanced compared with that from an AC case. This result was clearly linked by the authors to the precipitation response seen in the change in hardness levels (heat-treated versus as-built states). Wang et al. [4], also after working on L-DED of an Al-Sc alloy, recommended that attention should be paid to the integrated control of the solidification conditions and thermal cycling (thermal management), which are crucial for direct age-hardening in their opinion.

In terms of WAAM cases, Hou et al. [11] accomplished high-strength parts for an Al-Mg-Sc-Zr alloy, especially after the direct age-hardening of as-built samples produced with NC with an IT of 100 °C (varied dwell times for cooling) as a thermal management solution. The hardness/strength of the parts built with IT control was almost the same as for samples produced with a non-stop deposition strategy (without dwell times for cooling). But the hardness/strength of the first case was significantly higher after direct age-hardening, which indicates the importance of thermal management. These authors also detailed that the strengthening of Al-Mg-Sc-Zr alloys, which is the case of Scalmalloy^®, might take place via solid solution, precipitation, grain refinement, and dislocation mechanisms (indeed, as verified, exhibiting a decreasing participation in that order).

In another WAAM case, Ponomareva et al. [12] used mainly passive thermal management (continuous cooling of the building platform applied by a fan/blower) for an Al-Sc alloy, and a supersaturated solid solution with Sc was achieved after deposition, which was demonstrated as being prone to direct age-hardening. But in this case, the WAAM processing time was quite long due to deposition pauses between the layers for part cooling (significant aid of natural thermal management). The same authors also mentioned that in addition to greater alloying, the mechanical properties of the material could be potentially increased by means of the achievement of a higher cooling rate at deposition. This scenario could be prospectively implemented by active cooling approaches for thermal management, as they are generally more effective for heat dissipation from the part during WAAM [5].

The results from Xia et al. [13], who explored WAAM of an Al-Mg-Sc alloy with an IT of 80 °C and substrate preheating to 120 °C (via a heated building platform) as a thermal management solution, show that Sc (once in supersaturated solid solution) precipitated as Al₃(Sc,Zr) during deposition due to the WAAM thermal cycle. Their findings might likely relate to faster cooling rates right after deposition/solidification for leaving Sc into the solution at first and then to slower cooling rates and hence longer exposure times to higher/aging temperatures later (yet during the WAAM processing time) for the formation of precipitates. As the same authors explain, a transformation from columnar to equiaxed grains occurred in the deposits due to the presence of Sc, accompanied by a significant reduction in grain size. As they further argue, during sequential deposition, each layer will have a certain amount of metal remelted, with the other part (not remelted) heated like in heat treatment procedures.

Although Scalmalloy^® processing via WAAM has been shown to be feasible already, lack of proper thermal management can lead to heat accumulation [10] and hence, in this case, to the overaging of the parts being built [14] and, in turn, to declined mechanical performance. One might expect that Scalmalloy^® overaging could be even more detrimental in thin-walled structures given the typical relatively high heat input from the arc-based AM process and due to the impaired heat sinking through such a geometry [10]. The overaging problem resides in the fact that coarsened Sc aluminides cannot be redissolved and reprecipitated by the application of post-WAAM heat treatments because their solubilization temperature is close to the melting temperature of the Al-Mg matrix [1]. Moreover, although natural and passive thermal management solutions can help with material performance both after deposition and after heat treatment, they might hinder productivity due to the long cooling times typically utilized (required) between layers. And based on the insights from the related literature, in addition to potentially enhance productivity, the application of the NIAC technique (faster cooling rates and therefore shorter exposure times to higher/aging temperatures) could help in preventing the formation of the primary Al₃(Sc,Zr) phase during deposition, leaving Sc still in a solid solution for age-hardening later on (during post-WAAM heat-treatment).

Therefore, the present work aimed to assess the feasibility of processing thin-walled Scalmalloy^® (Al-Mg-Sc-Zr) structures by WAAM while examining the effects of powerful thermal management conditions (via arc energy level and heat dissipation intensity selection) on their response to direct age-hardening heat treatment (without solution annealing). As a complement, the geometry, porosity, and processing time of such parts were also analyzed.

2. Materials and Methods

Thin-walled Scalmalloy^® preforms were deposited under a technique called near-immersion active cooling (NIAC) as well as under natural cooling (NC) for thermal management by using the experimental rig shown in Figure 1. Under the NIAC technique, the preform is deposited inside a work tank that is filled with a cooling liquid (for forced convection), whose level rises as the metal layers are deposited. The distance separating it from the deposition level (layer edge to water distance—LEWD) is the main parameter for heat sinking control in this active thermal management technology [15]. The tank was simply kept without water for the NC approach (unforced air convection and radiation only—used for comparison basis). In this case, the procedure involved the wait of a cooling time to reach a specific interpass temperature (IT) before the onset of each next layer. Such a temperature value was estimated from measurements taken from deposits with the NIAC technique. Therefore, the IT level used for the NC approach was equivalent to the level found in an NIAC case.

The depositions were carried out on vertical substrates (Al5052 clamped plates) to mimic, since the first deposit/layer, the heat flux conditions and mechanical stiffness that typically dominate wall-like preforms. Each substrate was lightly sanded just before the first deposition and the wire surface was continuously cleaned by a piece of felt rubbing to it just before the action of the drive rolls of the wire feeder. An extended (large diameter) shielding gas nozzle was also employed as a way of mitigating the excessive oxidation of the surface of the layers, as successfully implemented in a previous work on the WAAM of Al alloy parts [16]. Other general deposition details are summarized in

Table 2. General deposition conditions.

Arc AM equipment	Fronius CMT 1—TransPuls Synergic 5000
Synergic line code	CMT 1070
Shielding gas	Commercially pure Ar at 12 L/min
Wire	Scalmalloy® with 1.0 mm in diameter
Substrate	Al5052 (280 mm × 38.1 mm × 6.35 mm)
CTWD 2	12 mm
Preform geometry	Thin walls with 50 mm in height 3 and 250 mm in length
Building strategy	Single-pass multi-layer bidirectional depositions
Cooling liquid for NIAC	Tap water (20–30 °C)
Dwell time for NIAC	5 s after/before each layer 4

¹ Cold metal transfer; ² contact tube to work distance; ³ the number of layers changed according to each deposition condition as needed to reach approximately 50 mm in height; ⁴ for NC, the dwell time varied to achieve the IT level selected.

.

Concerning the NIAC cases, two deposition conditions were applied, with one based on the variation in the wire feed speed/travel speed (WFS/TS) ratio (by changing the TS level) while keeping the LEWD parameter unchanged (

Table 3. Experiments with variation in the WTS/TS ratio for the same LEWD value.

Run	WFS (m/min)	TS (mm/s)	WFS/TS	Z Increment (mm)	LEWD (mm)
1	6.5	4.4	24.5	3.2	20
2		6.0	18.0	2.7
3		8.7	12.5	2.2
4		13.6	8.0	1.6
5		24.1	4.5	1.4

) and the other based on varying the LEWD parameter while keeping the WFS/TS ratio unchanged (

Table 4. Experiments with variation in the LEWD parameter for the same WTS/TS ratio.

Run	WFS (m/min)	TS (mm/s)	WFS/TS	Z Increment (mm)	LEWD (mm)
6	6.5	8.7	12.5	2.2	10
7 1					20
8					30
9					40
10					NC 2

¹ Run 7 is the same as run 3 of Table 3; ² the IT level in this case (for comparison basis) was the same as measured in run 3 of Table 3 (run 7).

). Thus, the arc energy per unit length of deposit (AE) level, as a consequence of changes in the WFS/TS ratio, was also taken as a thermal way of potentially affecting the direct age-hardening response of the material after WAAM. Accordingly, this approach can also be considered as thermal management (of intrinsic type, as it is implemented via deposition process parametrization) [5]. In the end, as also listed in Table 4, a thin-walled preform of reference was deposited under the NC approach with the IT level being estimated from the temperature measurements taken during run 3 of Table 3 (a preform deposited with the NIAC technique with a fixed LEWD value). The deposition parameters were selected from a work envelope previously developed for WAAM with an ER5356 wire (AlMg5) [16]. As a matter of fact, the chemical compositions found in ER5356 and Scalmalloy^® specifications are similar in terms of major elements and, by being so, a comparable behavior is expected in terms of deposition parametrization. The chemical composition of the Scalmalloy^® feedstock material utilized is shown in

Table 5. Scalmalloy^® chemical composition in weight % (according to the supplier).

Wire Spool	Mg	Sc	Zr	Ti	Al
1	4.68	0.42	0.086	0.13	Balance
2	4.64	0.44	0.095	0.15	Balance

.

Figure 2 shows the infrared (IR) pyrometers as arranged for constant measuring of the IT levels (trailing/leading IR pyrometer, labeled depending on the deposition direction of each sequential layer), as well as for surveying of the thermal history of a specific layer (fixed IR pyrometer). Data from a K-type thermocouple were used to adjust the emissivity of both these pyrometers to 22%. The measuring spot target of the trailing/leading IR pyrometer was pointed at 30 mm from the arc center at the layer since its signal became very noisy for shorter distances. The fixed IR pyrometer was set to constantly target the mid-length of the 5th layer of each preform. The specification of the IR pyrometers used is presented in

Table 6. Specification of the IR pyrometers used for monitoring of the preforms.

IR pyrometers	Mikron MI-PE140 with focusable optics
Spectral range	3–5 μm
Temperature range	30–1000 °C
Spot size	Ø of 2.9 mm at a distance of 380 mm
Emissivity	22%
Resolution	0.1 °C

.

The WAAM processing times, comprising the periods for the formation of the layers (actual deposition times) and those demanded between them for cooling the parts being built (idle times), were also documented for all the preforms.

After deposition, the thin-walled preforms were separated from the substrates and cut in half lengthwise by means of a bandsaw. Thus, one half of each case was submitted to direct age-hardening heat treatment at 325 °C for 4 h (without any previous solution annealing) followed by water quenching [17].

Scalmalloy^® phase formations are relatively well understood and diverse analytical techniques have been listed in the current literature to obtain detailed insights about the resultant microstructure, as per Kürnsteiner et al. [3]. Such authors employed scanning electron microscopy (SEM) with energy-dispersive X-ray spectroscopy (EDS) and field emission gun (FEG), transmission electron microscopy (TEM), and atom probe tomography (APT) to characterize the formation of the typical precipitates, but their occurrence was still verified via hardness testing (both at micro- and nano-scales). The authors indeed highlighted that the response to the presence of the precipitates is reflected in the mechanical properties of the material. In another example, Wang et al. [18] also used analytical techniques (electron back scattering diffraction (EBSD), X-ray diffraction, and TEM for characterizing the formation of the precipitates but relied on tensile testing for measuring their practical effects on material performance. Thus, as these analytical techniques generally involve high complexities/costs and are quite time-consuming, the present work employs a straightforward approach. This way, more industry-accessible methods were selected to be used here to evaluate the samples from a broader perspective, focusing on part features such as cross-section (analyzed via simple digital image processing), density (assessed via gravimetric buoyancy measurement), and hardness profiles (revealed through a digital Vickers microhardness tester).

Concerning geometrical features, the total wall width (TWW) and effective wall width (EWW) values of all preforms were determined through cross-section digital analyses (ImageJ software—version 1.54j) as illustrated in Figure 3. Then, the respective surface waviness (SW) values were calculated according to Equation (1) to quantify the resultant aspect quality. In addition, during the dwell times, each preform height was measured at mid-length locations along each layer by using a vernier caliper. These measures were then used to estimate the average layer thickness (LT) in each case.

$$\mathrm{S}\mathrm{W}={\displaystyle \frac{\mathrm{T}\mathrm{W}\mathrm{W}-\mathrm{E}\mathrm{W}\mathrm{W}}{2}}$$

(1)

The porosity level of samples with approximately 100 g taken from the preforms was estimated by using the gravimetric buoyancy method (Archimedes’ principle) with a digital precision scale (with 0.01 g of resolution) and distilled water. Next, the relative density (taken as a degree of porosity) of each case was estimated in relation to the density of a sample of the Scalmalloy^® feedstock material itself (considered as fully dense for reference), which was calculated via the same method. But, in reality, the porosity results must be interpreted in terms of total volume of voids, since potential internal cracks and/or lack of fusion defects also affect the results obtained from the Archimedes’ principle.

The microhardness profiles were surveyed along cross-section samples (one for each thin-walled preform), as illustrated in Figure 4. Such samples were first machined to obtain a flat surface and then finished according to standard procedures. A load of 1 kg (HV1) and a step distance of 1 mm were employed for the indentations. Then, the direct-aging response (DAR) index was estimated according to Equation (2).

$$\mathrm{D}\mathrm{A}\mathrm{R}={\displaystyle \frac{{\mathrm{H}}_{\mathrm{h}\mathrm{t}}-{\mathrm{H}}_{\mathrm{a}\mathrm{b}}}{{\mathrm{H}}_{\mathrm{a}\mathrm{b}}}}\ast 100$$

(2)

where

DAR: direct-aging response (%);

H_ht: average hardness in the heat-treated condition;

H_ab: average hardness in the as-built condition.

Figure 4. Hardness measurement path applied in the cross-sections of the preforms.

Figure 4. Hardness measurement path applied in the cross-sections of the preforms.

3. Results

The WFS/TS ratio and the AE level are both very important parameters for WAAM, as they define heat input (with all its thermal consequences) and affect layer geometry [11]. And for arc deposition processes with a consumable electrode, such as the CMT technology, these parameters operate in connection. For validation of that assumption in the present work, the corresponding AE values were plotted against the WFS/TS ratios selected, as shown in Figure 5. As observed, the relationship between the AE level and WFS/TS ratio was directly proportional and fitted a linear tendency. In practice, the confirmation of this result indicates, at least for the conditions presently applied, that the higher the WFS/TS ratio used, the higher the AE level operated by the deposition process.

3.1. Results on Geometrical Features, Porosity, and Processing Time

All thin-walled Scalmalloy^® preforms that were produced in this work presented shiny and smooth layer surfaces. And the resultant width remained virtually constant along the heights of such parts, suggesting no incidence of heat accumulation during the WAAM processing time in any case.

Figure 6 shows the lateral surface aspects and the respective cross-section views of the preforms deposited under the different WFS/TS ratios for the LEWD parameter fixed in 20 mm. And

Table 7. Resultant geometrical features (TWW, EWW, and SW), relative density (RD), processing time (PT = deposition times + idle times), and number of layers (NL) of the thin-walled Scalmalloy^® preforms built via WAAM with variation in the WFS/TS ratio (AE level) for the LEWD parameter fixed in 20 mm.

WFS/TS	TWW (mm)	EWW (mm)	SW (mm)	LT (mm)	RD (%)	PT (min)	NL
24.5	6.6	5.6	0.5	3.5	98.4	18.8	16
18.0	5.9	5.3	0.3	2.8	98.3	18.4	20
12.5	5.3	4.9	0.2	2.2	98.5	16.8	24
8.0	4.8	4.0	0.4	1.7	98.0	15.5	30
4.5	3.5	3.2	0.15	1.2	97.1	14.9	44

compiles the values found for the geometrical features analyzed and presents other details of the parts produced. As expected, for the same WFS level, the higher the TS value (the lower the AE level), the smaller the total wall width (TWW) and the smaller the average layer thickness (LT). The consequent surface waviness (SW) values tended to be lower the smaller the WFS/TS ratio selected. In addition, the relative density (RD) levels remained high (close to the Scalmalloy^® feedstock material, considered fully dense) and almost unchanged throughout all cases, suggesting a low degree of porosity in the preforms produced. Likewise, the WAAM production time (PT) results were overall similar, indicating that the period expended to build the parts was only marginally affected by the number of layers (NL) required to achieve the same wall height in each case (about 50 mm), which in turn depended on the LT results.

As shown in Figure 7 and compiled in

Table 8. Resultant geometrical features (TWW, EWW, and SW), relative density (RD), processing time (PT = deposition times + idle times), and number of layers (NL) of the thin-walled Scalmalloy^® preforms built via WAAM with variation in the LEWD parameter for the WFS/TS ratio fixed in 12.5 (same AE level) (NC case included for comparison basis).

LEWD (mm)	TWW (mm)	EWW (mm)	SW (mm)	LT (mm)	RD (%)	PT (min)	NL
10	5.2	4.8	0.2	2.25	98.1	16.8	24
20	5.3	5.12	0.2		98.5
30	5.4	5.12	0.15		98.1
40	5.2	4.9	0.15		96.1
NC 1	5.5	5.1	0.2		96.2	121

¹ The IT level for this case (for comparison basis) was kept between 65–75 °C.

, for the WFST/TS ratio fixed in 12.5 (same AE level), the variation in the LEWD parameter of the NIAC technique had no significant effect on the geometrical features of the preforms. In the case of the NC thermal management condition, as expected, if the interpass temperature (IT) level is equivalent to that obtained with the NIAC technique, the geometry of the thin-walled preform tends to be similar, but the PT duration was significantly increased (more than seven times longer) due to the cooling times demanded to reach the same thermal state between the layers. This indicates a remarkable loss in time-based productivity despite the fact that the same number of layers (NL) with the same vertical dimension (LT) were required to achieve the same wall height (about 50 mm) in both cases. In addition, the RD levels remained reasonably high and almost unaffected in all cases; again, a sign of a low degree of porosity in the parts produced.

3.2. Results on Thermal Features

3.2.1. Interpass Temperature Estimation

As described in previous work, the temperature data from the trailing/leading IR pyrometer during all the WAAM processing time tend to fit a bimodal distribution, and the first quartile (Q1) is representative of the average interpass temperature (IT) level along the layers [10]. Accordingly, Figure 8 shows the boxplot of the temperature data measured by the trailing/leading IR pyrometer in the Scalmalloy^® preforms produced under the NIAC technique, where Q1 is the bottom line of the box. As expected, for the same LEWD parameter, the lower the WFS/TS ratio (and AE level), the lower the average IT level. Conversely, for the same WSF/TS ratio (and AE level), the higher the LEWD parameter (the more distant the cooling water lamina is from the layer being deposited), the higher the average IT level (a direct consequence of the NIAC weaker effect).

Table 9. Average interpass temperature (IT) according to the WFS/TS ratio (and AE level) variation (fixed LEWD parameter = 20 mm).

	WFS/TS 24.5	WFS/TS 18	WFS/TS 12.5	WFS/TS 8	WFS/TS 4.5
IT (°C)	104	108	67	33	39

and

Table 10. Average interpass temperature (IT) according to the LEWD parameter variation (fixed WFS/TS ratio = 12.5 (same AE level)).

	LEWD 10 mm	LEWD 20 mm	LEWD 25 mm	LEWD 30 mm	LEWD 40 mm
IT (°C)	59	67	101	114	124

list the respective average IT values registered in such cases.

3.2.2. Cooling Rate Estimation

The typical thermal history (measured by the fixed IR pyrometer) experienced by a given layer of the thin-walled Scalmalloy^® preforms built via WAAM with the NIAC technique for active thermal management is presented in Figure 9. Each thermal cycle observed corresponds to the change in temperature due to the deposition of a new layer. As seen, the peak temperature level measured at a fixed point tends to decrease as the layers are sequentially deposited because the heat source (arc) moves away from the stationary measurement point. Unlike what is expected, the peak temperature values in the first two layers were lower than in the following ones. Such an occurrence may be related to arc interferences while it stays close to the measurement point. Therefore, the first two measurements were not considered for an estimation of the cooling rates in each preform, as presented next.

The peak temperatures observed in the thermal histories (as in Figure 9), except for the first two events, were then used to determine the peak temperature trends as a function of the distance from the top (last) layer for each preform, as exemplified in Figure 10. As noticed, if the distance from the top is extrapolated to zero (i.e., to the molten pool), the estimated peak temperature falls within the typical liquidus and solidus boundaries verified for Scalmalloy^®, which seems to be physically reasonable. This way, the cooling rate in each thermal cycle was estimated considering the temperature range between 90 and 70% of the respective peak temperature. An example of the correlation between the cooling rate and peak temperature values is shown in Figure 11. As presumed, the cooling rate tends to increase as the peak temperature level increases.

3.2.3. Resultant Cooling Rates

The effect of the WFS/TS ratio (and AE level) and LEWD parameter variations on the resultant cooling rate are respectively presented in Figure 12 and Figure 13 (considering Scalmalloy^® layers with peak temperatures of 650 °C). As perceived, the higher the WFS/TS ratio (and AE level) and the LEWD value (and the IT level), the slower the cooling rate, which is in accordance with what has been postulated for heat transfer in arc welding [10] and, by extension, to the WAAM scenario. It is interesting to notice in Figure 13 the fact that the higher the LEWD value, the more the cooling rate tends to approximate to the estimation made for the NC thermal management approach, which will be further commented on later.

3.3. Results on Direct-Aging Response

From the generally flat typical hardness profiles exemplified in Figure 14, it can be presumed that Al₃(Sc,Zr) precipitation (typically evidenced by rises in hardness levels) did not take place (at least not in an extensive amount) during the WAAM processing time of the Scalmalloy^® preforms, no matter the thermal management condition applied. But this result was not expected for the NC approach in particular, since it differs from what has been reported by Rometsch et al. [14]. Nevertheless, the formation of this hardening phase could have possibly occurred for longer deposition times under higher heat inputs (higher WFS/TS ratios (and AE levels)) than those utilized in the present work.

Concerning the change in the average hardness levels after post-WAAM heat treatment, as exemplified in Figure 14 and then comprehensively presented in Figure 15, for all the thermal management conditions, the direct-aging response (DAR) index was always positive, i.e., the hardness increased significantly due to the direct age-hardening procedure applied, as depicted in Figure 16.

As clearly seen in Figure 16, the faster the cooling rate (resulting from the thermal management conditions applied), the greater the DAR index, which corroborates the results presented previously by Taendl et al. [19]. And it is worth remarking that the related increase in the average hardness levels (likely due to precipitation) in the preform built under the NC approach was comparable to those rises verified in the parts produced with the NIAC technique.

4. Discussions

4.1. Discussions on Geometrical Features, Porosity, and Processing Time

Already established in the literature [20], the effect of the WAAM parameters on the geometry of the thin-walled preforms followed the expected behavior that has been reported elsewhere. For a fixed LEWD parameter, the higher the WFS/TS ratio (and the AE level), the more irregular the geometry of the preforms produced (Figure 6), as indeed pointed out by the SW results (Table 7). The higher IT levels observed (Table 9) and the likely larger molten pool volumes formed with greater thermal charges (as seen in Figure 6 and indicated by the larger LT results listed in Table 7) explain these results, since in such conditions, the material tends to collapse laterally before solidification during deposition. This confirms the importance of starting to resolve thermal management concerns by means of the proper intrinsic parametrization of the deposition process to improve the possibility of efficient material usage (lower SW levels). The variation in the NIAC technique (by means of its LEWD parameter) for active thermal management did not change the geometrical results in a significant way in comparison with the NC case (as seen in Figure 7 and listed in Table 8). It is reasonable to think that layer geometry formation, given conditions without heat accumulation (as indeed provided by the application of the NIAC technique as well as by the NC approach with IT control), will be more linked to the deposition process parameters and to the immediate effects of the consequent heat input and metal transfer scenario (as confirmed by the effects of the WFS/TS ratio and AE level), such as on the molten pool size and arc pressure (if too large/high, these factors will play against stable layer formation).

The relative density levels of the as-built Scalmalloy^® parts produced with the NIAC and NC approaches (Table 7 and Table 8) were considered low (less than 4%) and within the range previously observed in the WAAM of a more common Al alloy [10]. This indicates that suitable Scalmalloy^® preforms (from the density perspective) can be produced via WAAM with the thermal management conditions applied, including with the NIAC technique with its water lamina level close to the deposition level for more heat dissipation capability. This is particularly important from a productivity point of view, since the NIAC technique can remarkably reduce the WAAM processing time (more than seven times shorter in the present case) in relation to an equivalent thermal condition (same IT level) under the NC approach for thermal management, as listed in Table 8.

And in addition to these expected yet significant results, another important finding, specifically regarding the geometrical features, was that the WAAM working envelope as developed by da Silva et al. [16] for an ER5356 wire was proven to be highly effective for the case of the Scalmalloy^® employed in the present work as well. Thus, future research on parameterization for the WAAM of Scalmalloy^® parts could prospectively initiate with a focus on an ER5356 feedstock material (given its relatively lower cost and more prompt availability) and subsequently proceed with the transfer of parameters to the material of actual interest.

4.2. Discussions on Thermal Features

From arc welding to arc AM technologies, it is well known that the interpass temperature (IT) is a fundamental parameter to control the cooling rate in the part being produced [21]. As indeed previously reported for the WAAM of Al parts, for a given deposition condition, regardless of the type of thermal management approach applied, the use of the same IT level resulted in similar wall widths, but the processing time and thermal history of the respective parts were significantly different [10].

Thus, in the present work, an approach was developed to estimate the cooling rate at the molten pool by utilizing the thermal history measured with a fixed IR pyrometer (Figure 10). The cooling rates that were estimated, ranging from 35 to 140 °C/s, align reasonably well with the values measured from 500 to 300 °C by Gierth et al. [20] for the WAAM of wall-like preforms made from AlMg5Mn via the CMT deposition process. For the changes in the NIAC technique intensity in the present work, with increasing LEWD values, the cooling rates tended to line up with that of the NC case (35 °C/s). As a matter of fact, this tendency sounds physically reasonable given the NIAC technique loses its power as the cooling liquid (water, in this case) lamina is selected to operate more distant from the deposition (arc) level (LEWD parameter increase).

From the results obtained in the present work, Figure 17 highlights the potential of the NIAC technique to modulate the cooling rate for a target wall width. For an EWW value aimed at 5 mm, for example, the cooling rate of the part being produced (layers being deposited) could be adjusted between 42 and 130 °C/s by adjusting the LEWD parameter. Thus, this NIAC feature could be exploited to allow for the WAAM of components made of Scalmalloy^® (or even of other metals) with more control over microstructure formation and hence with tailored mechanical properties/performance.

As found by Ponomareva et al. [12], the strengthening of Al-Sc alloys occurs because of artificial aging by the formation of nanosized Al₃(Sc,Zr) precipitates from the decomposition of the supersaturated Sc solid solution once formed at sufficiently high (>100 °C/s) crystallization rates. In fact, in the present work, the best result for Scalmalloy^® strengthening after age-hardening treatment took place overall for a faster cooling rate case during the WAAM processing time due to the application of the intrinsic and active thermal management solution. And this result is also in total agreement with Ren et al. [22], who highlighted, after a study on the WAAM of an Al-Sc alloy, that the higher the cooling rate of the melt, the more Sc remains in solid solution, which provides favorable conditions for the precipitation of a large amount of secondary Al₃(Sc,Zr) phase in subsequent thermal processing (heat treatment).

4.3. Discussions on Direct-Aging Response

During the solidification of Al-Sc-Zr alloys, part of the Sc content precipitates as primary Al₃(Sc,Zr) formations, refining the grains, and the rest remains in solid solution [1]. Such primary precipitates grow at the expense of Sc in solid solution, reducing the number of precipitates formed in solid state during direct age-hardening treatment. And the amount of primary Al₃(Sc,Zr) formation declines with an increase in the cooling rates experienced by the material under processing. At the same level of Sc content, the improvement of mechanical properties due to the precipitation of Al₃(Sc,Zr) in the solid state is more pronounced than that attributed to the grain refinement caused by primary precipitates formed during solidification [23]. Thus, it is always desirable to have the maximum amount of Sc retained in a supersaturated solid solution straight after the WAAM period.

Therefore, it could be considered that the ideal scenario for the WAAM of Scalmalloy^® would be the one in which the thermal management conditions (heat input and heat dissipation) resulted in such a fast-cooling rate that the maximum amount of Sc could be retained in a supersaturated solid solution and that the solid-state precipitation could be then avoided during the whole AM processing time. This way, Al₃(Sc,Zr) precipitation within a solid state can be guaranteed to be triggered significantly and stably by a post-WAAM heat treatment (direct age-hardening—without any need for solution annealing). Moreover, direct age-hardening and stress-relieving procedures could be potentially implemented in combination in a single step.

To assess the effect of thermal management on the direct-aging response of the thin-walled Scalmalloy^® structures produced via WAAM, the following hypotheses were formulated based on possible hardness behaviors:

(a)

Variation in the hardness profile along the preform height: If a significant hardness level variation is observed along the preform height, it indicates that primary Al₃(Sc,Zr) precipitation may have occurred to a varying extent during deposition due to the thermal history imposed by the WAAM process (the re-heating cycles and possible heat accumulation could change the hardness levels after each layer);

(b)

Positive direct-aging response: If the hardness level increases after direct age-hardening, it suggests that Sc may have remained in the supersaturated solid solution during deposition and that secondary Al₃(Sc,Zr) precipitation was likely activated by the post-WAAM heat treatment. In this case, the direct-aging response result should be proportional, among other factors, to the fraction of Sc that remained in the supersaturated solid solution;

(c)

No direct-aging response: If the hardness level remains largely unchanged after the post-WAAM heat treatment, it indicates that the part may have been already exposed to overaging in the as-built state and therefore could not respond to the direct age-hardening procedure applied;

(d)

Negative direct-aging response: If the hardness level decreases after direct age-hardening, it implies that primary Al₃(Sc,Zr) precipitation may have extensively occurred during the sequential depositions (likely not leaving any significant Sc content in the supersaturated solid solution), which could be driven to overaging by the post-WAAM heat treatment.

Based on the flat hardness profiles revealed along the height of all the thin-walled Scalmalloy^® preforms produced with the thermal management conditions applied (exemplified in Figure 14) and on the always-positive direct-aging response (DAR) results observed (increase in hardness seen Figure 15), it can be inferred that hypothesis b is more plausible and, as such, a large fraction of the Sc content must have indeed remained in the supersaturated solid solution during the WAAM processing time. And that seems to have been the case, not only with the NIAC application but even with the NC approach (likely due to the reasonable low IT level employed). Thus, Al₃(Sc,Zr) precipitation must have occurred predominantly during the post-WAAM heat treatment (direct age-hardening) for all the thermal management conditions presently examined. And the hardness levels observed in the as-built as well as in the heat-treated state in the Scalmalloy^® parts produced in this work were comparable to those reported in the related current literature (listed in

Table 11. Hardness reported in the current literature and as found in the present work for different Al-Sc alloys with the respective manufacturing processes, thermal management approaches, and heat treatment conditions applied, as well as resulting direct-aging response (DAR) index values.

Ref.	Chemical Composition (Weight %)	Manufacturing Process	Thermal Management	Age-Hardening Treatment		Hardness (As-Built)	Hardness (Heat-Treated)	DAR (%)
[4]	Al-4.86Mg-0.5Sc-0.21Zr-068Mn-0.20Si	L-DED	Passive	325 °C	4 h	92 HV0.3	116 HV0.3	26
[11]	Al-6.54Mg-0.36Sc-0.11Zr	WAAM (Arc-DED)	Natural	325 °C	6 h	87 HV1	118 HV1	36
[12]	Al-5.82Mg-0.42Mn-0.19Sc-0.23Cr	WAAM (Arc-DED)	Passive and Natural	300 °C	6 h	89 HV1	118 HV1	33
[14]	Al-0.9Sc	L-DED	-	325 °C	2 h	34 HV5	65 HV5	48
[19]	Al-4Mg-0.4Sc-0.12Zr	EB-Remelting 1	-	325 °C	3 h	64 HV0.1	110 HV0.1	42
[23]	Al-0.3Sc-0.15Zr	Casting	-	300 °C	3 h	44 HV10	71 HV10	38
[24]	Al-3.4Mg-1.08Sc-0.23Zr-0.5Mn-0.5Cu	L-PBF	Passive	300 °C	12 h	110 HV1	165 HV1	33
Present work	Al-4.7Mg-0.43Sc-0.09Zr-0.14Ti 2	WAAM (Arc-DED)	Intrinsic and Active (NIAC)	325 °C	4 h	94 HV1 3	126 HV1 3	34 3

¹ Electron beam (EB) process; ² mean composition of the wires employed in this work; ³ values of the highest DAR index level achieved in this work (run 3/7). Note: the average DAR index value found in the literature was 36.6% for all processes identified and specifically 34.5% for other WAAM cases.

) and therefore resulted in similar DAR index values (up to 34% in the present work against an average of 36.6% for all processes found in the literature and specifically 34.5% for other WAAM cases).

From the correlation between the WTS/TS ratios (and AE levels) with the respective cooling rates applied and the resultant age-hardening results (Figure 16a), it was found that the latest feature of the Scalmalloy^®, as explored, reduced with increasing WFS/TS ratios (and AE levels) and consequent decreasing cooling rates. This behavior can be explained by the fact that greater thermal charges due to higher AE levels (higher WFS/TS ratios) resulted in slower cooling rates. And a reduced cooling rate certainly promoted primary precipitation, decreasing the availability of Sc for precipitation in the solid state later on during the post-WAAM heat treatment (direct age-hardening). According to Ren et al. [22], from a study about the effect of Sc content on the microstructure and properties of an Al-Sc alloy processed via WAAM, the amount of secondary Al₃Sc is directly related to the amount of Sc in the supersaturated solid solution in the Al matrix, which, in turn, is directly linked to the cooling rate of the melt. The results of their work revealed that the heat treatment applied (direct age-hardening at 350 °C for 1 h) had two effects on the precipitated phases. First, the authors considered that it provided energy to the phases precipitated at the grain boundaries so that the elements diffused into the grains and precipitated within them, thus refining the grain boundaries and evenly distributing the precipitates. Second, they considered that it promoted the precipitation of the secondary Al₃(Sc,Zr) phase. Similar effects must have been responsible for the significantly positive change in hardness (after direct age-hardening) of the Scalmalloy^® samples produced via WAAM in the present work based on the use of an intrinsic (arc energy selection) and active (NIAC technique application) thermal management solution.

As shown in Figure 16b, for the WFS/TS ratio set in 12.5, the variation in the LEWD parameter of the NIAC technique had low impact on the age-hardening results. This almost absence of effects may be explained by the fact that, under such conditions, the DAR index was likely already close to the maximum for the Scalmalloy^® feedstock material employed. The influence of the LEWD variation over the age-hardening results could have been certainly more evident for a WFS/TS ratio of 18 or 24.5 (see Figure 16a) and hence with higher AE (and heat input) levels, a scenario for which the powerful heat sinking capability of the NIAC technique really becomes an effective asset.

Lastly, it is worth highlighting the fact that the NIAC technique for active thermal management also permitted the acceleration (due to faster cooling rates and hence shorter WAAM processing times) of the achievement of DAR index levels similar to the one accomplished with the application of the NC approach under an equivalent general thermal condition (same IT) (Table 8). If compared with the related current literature, Ponomareva et al. [12], for instance, found a DAR of 33% (about the same as achieved in the present work) but, in that case, the WAAM processing time of the Al-Sc alloy part was quite long due to an average of 60 s of pause for cooling between the layers via a passive (reduction in the substrate temperature to below 100 °C after each layer by the building platform continuous cooling applied by a fan/blower) and natural thermal management hybrid solution—against only 5 s in the present work with the NIAC technique (IT of 33–108 °C). Thus, the result found here corroborates the productivity advantage that the application of the NIAC technique has demonstrated for other Al alloys [10].

4.4. Industrial Impact and Practical Applications

This work provides valuable insights into the WAAM of Scalmalloy^® (Al-Mg-Sc-Zr), particularly regarding how a more effective thermal management solution influences its response to direct age-hardening. And given the growing industrial interest in WAAM for producing high-strength and low-weight structural parts, specifically from distinctive aluminum alloys for aerospace applications, the findings reported here are both opportune and relevant.

A key contribution, with potential for direct impact in related practical applications, is the comparative evaluation between near-immersion active cooling (NIAC) and natural cooling (NC) as strategies for thermal management in WAAM, which are shown to have a significant impact on productivity. As summarized in

Table 12. Competitive edge for applying NIAC and NC as strategies for thermal management in the WAAM of Scalmalloy^® parts.

Main Features Affected	Near-Immersion Active Cooling (In Water)	Natural Cooling (In Air)
Geometrical quality	Excellent	Excellent
Porosity level	Low	Low
Productivity for small–medium parts	High (continuous deposition)	Low (need of cooling intervals)
Productivity for medium–large parts	High (continuous deposition)	High (continuous deposition)
Direct-aging response	High	High
Key advantage for application	High productivity with the ability to tailor microstructure and mechanical properties	Easy implementation on any equipment, especially robot cells with positioners (moving tables)

, while both these approaches for thermal management can result in parts with excellent geometrical quality and low porosity levels, the NIAC technique enables more productivity, especially for small-to-medium-sized parts for allowing continuous deposition without heat accumulation concerns (large-sized parts tend to facilitate heat dissipation). In contrast, the use of NC is more adaptable to robotic systems and positioners. Additionally, and not least important given the context of the present work, both thermal management solutions can support the effective responses of Scalmalloy^® parts produced by WAAM to direct age-hardening.

Thus, from a practical perspective, these insights might help optimize WAAM processing conditions for different industrial scenarios. If the priority is the rapid manufacturing of smaller parts, the NIAC application is particularly advantageous due to overall production time reduction and part quality preservation. Meanwhile, NC is better suited for flexible manufacturing environments, such as WAAM cells with rotating positioners, making this simpler thermal management approach indicated not only for larger but also for more complex structures.

5. Conclusions

This work essentially aimed to assess the effects of thermal management conditions on the age-hardening response of thin-walled structures made from a Scalmalloy^® (Al-Mg-Sc-Zr) feedstock material via WAAM, which were in due course built both with high geometrical quality (low surface waviness) and low porosity levels. Overall, from hardness testing results, it can be inferred that the direct-aging response of the material after the WAAM period (via heat treatment) depended mainly on the WFS/TS ratio variation (selection of the arc energy per unit length of deposit—intrinsic thermal management). In contrast, other thermal management approaches (of active or natural type) likely helped in maintaining Sc in the supersaturated solid solution during deposition and, by doing so, assisted in the formation of precipitates during post-WAAM heat treatment. Thus, Scalmalloy^® strengthening was demonstrated as feasibly triggered by means of a post-WAAM direct age-hardening treatment solely (without previous solution annealing). Additionally, in comparison with a thermally-equivalent (same interpass temperature) condition based on natural cooling (NC), the near-immersion active cooling (NIAC) technique allowed the achievement of such a positive result on direct-aging response with remarkable improvement in productivity in view of the fact that the WAAM processing times were significantly shortened. In terms of the numbers then accomplished for the Scalmalloy^® parts built via WAAM, the main conclusions can be summarized as follows:

➣

The geometrical aspect of the parts was positive and similar despite the changes in thermal management, with the surface waviness ranging from 0.15 to 0.5 mm;

➣

The porosity of the parts was not an issue and similar despite the changes in thermal management, with the relative density ranging from 96.1 to 98.5% (considering the Scalmalloy^® feedstock material fully dense as reference);

➣

The WAAM processing time could be reduced down to 16.8 min with the application of the NIAC technique against 121 min needed with the NC approach (an impressive 86% reduction);

➣

The response to direct age-hardening was always positive, with the highest level of increase in hardness at 34% as a consequence of the fast cooling rates provided by the intrinsic (arc energy selection) and active (NIAC technique application) thermal management solution employed.

As subsequent steps for related investigations, the findings of the present work, at least for the best case of the NIAC application, as well as for the NC case for a reference basis, should be next deepened to find and better understand ways for the optimum parametrization of the WAAM process (for intrinsic thermal management) and of the NIAC technique (for active thermal management). Hence, assessing mechanical properties (other than hardness) of Scalmalloy^® parts produced as in this work and conducting an in-depth analysis of the microstructural strengthening mechanisms involved via advanced analytical methods, as well as proceeding with an investigation to separate their contribution to mechanical performance, both in the as-built and heat-treated states, for example, should be considered.