Influence of Stress Relief Annealing Parameters on Mechanical Properties and Decomposition of Eutectic Si Network of L-PBF Additive Manufactured Alloy AlSi10Mg

Abstract

This paper evaluates the effect of stress-relieving heat treatment on the AlSi10Mg alloy prepared by additive manufacturing using the Laser Powder Bed Fusion (L-PBF) with print parameters: 370 W, 1400 m/s, and 50 μm. The as-built state and four different annealing modes (240 °C/2 h, 240 °C/6 h, 300 °C/2 h, and 300 °C/2 h/water-quenched) are investigated. To determine the effect of the annealing mode on the mechanical properties of the L-PBF AlSi10Mg alloy, heat-treated samples were compared with the as-built state and with each other. The mechanical properties of the samples were determined by tensile and hardness tests. The strength in the as-built state is 488 MPa, depending on the method of heat treatment, the strength values range from 296 MPa to 417 MPa, and the HV10 hardness values are in accordance with the measured strength values. Furthermore, the microstructure of the samples was investigated by scanning electron microscopy (SEM) analysis, which was then linked to the measured mechanical properties. The composition of the microstructure of the alloy and its influence on the mechanical properties were determined by energy dispersive spectroscopy (EDS) analysis. Furthermore, the differences between the individual heat treatments in comparison with the as-built state were analyzed and the phenomenon of decomposition of the silicon network after reaching specific temperatures was discussed and confirmed. The paper evaluates the effect of dwelling time on stress relief annealing. It was found that if annealing at intermediate temperatures of 240 and 300 °C is applied, changes in structure and mechanical properties are more temperature- than dwell-time-dependent.

1. Introduction

Aluminum alloys are the most common and widely used non-ferrous casting materials [1]. High-quality castings with suitable mechanical properties are produced by high-pressure casting, which is considered the best casting method for aluminum alloys among the available casting methods [2,3,4]. Specific casting structure and properties are well described [5]. In contrast, aluminum alloy components produced by 3D printing using L-PBF technology have a completely different microstructure than cast alloys and thus different properties [2]. By 3D printing these components, comparable or better mechanical properties can be achieved than by high-pressure casting technology [6]. The heat treatment process and its effect on mechanical properties and microstructure is also well known and described for aluminum casting alloys. For casted aluminum alloys, typical heat treatment (HT) is solution annealing followed by artificial aging (T6) [2,7], which generally leads to an increase in strength properties compared to the as-cast state. In contrast, it was found that the use of T6 heat treatment in the case of the 3D printed AlSi10Mg alloy leads to a significant decrease in strength characteristics compared to the as-built state; this fact is given by a different structure of the Laser Beam Powder Bed Fusion (L-PBF) parts in the comparison with casted ones [7,8,9]. For components after 3D printing, stress relieving is usually used [7,10], but it is not yet entirely clear which parameters of mentioned heat treatment are suitable for the final application of the material thus prepared and processed.

The L-PBF process is based on the melting of small volumes of powder by a high-energy source, in this case, a laser. Unlike casting, L-PBF is characterized by an extremely high solidification rate which has been reported to be of the order of 10³–10⁵ K/s [11,12]. The layered structure of the printed part when the melting pools overlap brings another phenomenon—cycles of reheating and cooling of already solidified material caused by printing of adjacent or overlapping layers. These facts have a significant effect on the metallurgy of the material, which among other things has the following impacts:

• Supersaturated solid solutions and residual stresses are formed and occur in the alloy during the printing, primarily because of the high cooling rate.
• Heat treatment of previously solidified material subsequently causes local phase transformations and precipitation.

Heat treatment of L-PBF aluminum alloys should be considered as a separate issue. From the differences discussed above, it follows that the same heat treatment procedures cannot be applied to printed alloys as for cast ones expecting the same results in the change in microstructure and the associated change in mechanical properties.

The currently used HT of L-PBF additive manufactured AlSi10Mg can be divided into three categories according to the used temperature, specifically, low-, intermediate-, and high-temperature annealing, each with a different effect on the final microstructure. Annealing at low temperatures can be classified as an HT up to 200 °C equal to direct aging (T5). Stress relief annealing as an intermediate-temperature HT up to 350 °C, and high-temperature HT as solution annealing up to 550 °C, which can be also followed by direct aging (T5).

HT is applied to (i) relieve stress caused by the nature of the L-PBF production process, as previously discussed mainly by the high cooling rate and cyclic reheating of already solidified material, and (ii) to alter the mechanical properties of the alloy.

This article focuses on intermediate-temperature annealing. This research aims to determine how the heat treatment parameters, such as temperature, dwell time, and cooling rate, influence the mechanical properties and microstructure of the L-PBF AlSi10Mg alloy. HT parameters recommended by a powder manufacturer [13] and the studied literature were used and modified to develop a comprehensive view of the issue of annealing at intermediate temperatures.

This work evaluates modes of intermediate-temperature heat treatment (stress relieving), within one build cycle, to better compare the direct effect of HT on the as-built state and its properties and microstructure. In the available literature are papers that evaluate mainly HT at 300 °C/2 h in various combinations, such as a comparison with low- and high-temperature annealing [7], low-temperature HT, and hot isostatic pressing (HIP) [14], or after stress relieving (SR) compared to SR + HIP [15]. Another evaluated HT is 300 °C/2 h/water-quenched, and in these papers, the HT is usually evaluated in comparison with solution annealing (high-temperature HT) [16] or also followed by artificial aging (equal to T6) [17]. Several papers only evaluate intermediate-temperature HT, namely Tang and Pistorius [10], who evaluate only the mode of 300 °C/2 h and its effect on porosity oxides and fatigue resistance. Wang et al. [18] studied 300 °C compared to 270 °C and 330 °C and the effect of building orientation and heat treatment on the anisotropic tensile properties, and Rosenthal et al. [19] compared 300 °C to the manufacturer’s recommendation and gravity and high-pressure casting. Patakham et al. [20] evaluated melt pool boundary characteristics and Si morphologies on mechanical properties and fracture behavior after HT 270 °C/1.5 h.

This paper studies (a) 240 °C/6 h as the manufacturer’s recommendation [13], a modified version (b) 240 °C/2 h in order to find out whether the annealing time can be shortened to make the process more efficient, and further (c) 300 °C/2 h and (d) 300 °C/2 h/WQ and their influence on mechanical and other properties. The as-built state is used to study the influence of individual HT modes. Although the effect of HT on the structure and properties of the L-PBF AlSi10Mg alloy has been previously studied, this specific combination of printing parameters and HT modes, and specifically the modes of HT 240 °C/6 h and 240 °C/2 h, have not been studied in the available literature.

2. Materials and Methods

AlSi10Mg powder CL 31AL by Concept Laser GmbH with the nominal composition given in the following

Table 1. Chemical composition of AlSi10Mg powder by Concept Laser GmbH in wt. % [13].

Si	Mg	Fe	Mn	Ti	Cu	Zn	C	Ni	Pb	Sn	Al
9.0–11.0	0.20–0.45	0–0.55	0–0.45	0–0.15	0–0.10	0–0.10	0–0.05	0–0.05	0–0.05	0–0.05	Bal.

was used.

The manufacturer guarantees a particle size distribution in the material sheet: 99.9% up to 45 µm, 89.1% up to 38 µm. SEM image of the used powder is in Figure 1. In terms of size distribution, the powder is highly heterogeneous. The scanning electron microscopy (SEM) analysis in Figure 1 shows that considerably larger particles (up to 70 µm) are present in the powder than guaranteed by the manufacturer. It can be a combination of several powder grains melted together, or an incorrectly set up production process. The imperfect spheroidization of the powder particles is also clear from the figure. All these facts, in combination with the printing parameters, can affect the resulting mechanical properties of the printed material. However, the occurrence of elements controlled by energy-dispersive spectroscopy (EDS) analysis corresponds to the chemical composition given by the material sheet.

For the additive manufacture of AlSi10Mg L-PBF technology was used. Samples were printed in a Concept Laser M2 machine in the Z axis in a protective nitrogen atmosphere with the printing parameters given in

Table 2. Printing parameters used for AlSi10Mg L-PBF samples.

	Power (W)	Speed (mm/s)	Layer Thickness (µm)	Spot Size (µm)
Skin	200	800	25	190
Core	370	1400	50	190
Support	200	1600	50	50

. The parameters that were chosen are the optimal and used settings for the given printer and powder based on previous experience at the given workplace.

The samples were divided into groups of 5 and heat treated at several temperatures; 5 samples were left in an as-built state. The individual modes of heat treatment and the method of cooling are listed in

Table 3. List of heat treatment modes of L-PBF AlSi10Mg used in the experiment.

Identification of the Sample Group	Description of the Heat Treatment
AB (as-built)	With no further heat treatment after printing
T	240 °C/6 h; FC to 100 °C then AC
Z	240 °C/2 h; FC to 100 °C then AC
E	300 °C/2 h; WQ
P	300 °C/2 h; FC to 100 °C then AC

. All annealing temperatures were reached in 1 h, after dwelling at temperature cooling in different media followed, including furnace (furnace-cooled, FC), air (air-cooled, AC), and water (water-quenched, WQ).

All samples were machined at Slovtos S 280 CNC to the precise shape for the tensile test (short test rod: d = 8 mm, d₁ = M12, L_o = 40 mm, L_t = 77 mm) according to the standard (ISO 6892-1).

A tensile test was performed on all samples according to the ISO 6892-1 standard on the universal testing machine INSTRON 5582 with a maximum load force of 100 kN and the applied crossbar speed of 5 mm/min.

Samples for metallographic analysis in the cross-section and longitudinal section were cut from the threaded heads of the tensile test specimens.

To reveal the microstructure, the samples were prepared through a standard metallography procedure (including mechanical grinding and polishing) and etched in Keller’s reagent (2.5% HNO₃; 1.5% HCl; 1.0% HF; 95.0% distilled water).

Metallographic analysis was performed on the samples using optical microscope Carl Zeiss Jena Neophot 32 and scanning electron microscope JEOL JSM-7600F) and subsequently the analysis of chemical composition by EDS using the Oxford X-Max detector 50 mm².

3. Results and Discussion

3.1. Mechanical Properties and Microstructure of the As-Built State

Although the structure of the L-PBF AlSi10Mg alloy in the as-built state has already been described in the available literature, it should be noted that the resulting structure is affected by several parameters (powder, printing parameters, printing direction, etc.). For this reason, analyzing the structure and properties in the as-built state for each experiment is always necessary.

Several phenomena cause the high strength of the as-built AlSi10Mg alloy:

• Hall–Petch reinforcement—grain boundary strengthening pile-up of dislocations at the cell boundaries (fine Si network) which further inhibits dislocation motion and considerably contributes to the high strength of as-built AlSi10Mg [21,22].
• Orowan strengthening in the as-built condition upon the dislocations is pinned by the eutectic Si phase at cell boundaries, the bend around eutectic Al phases, and the Orowan looping mode is activated. Dislocations glide through the cell boundary, thus highly strengthening the alloy. A large number of new dislocations are initiated by this mechanism, which increases the alloy’s strain hardening ability [8,23].
• Strengthening induced by localized shear stress: nano-sized Si particles and Si precipitates in the matrix [24].
• Dislocation strengthening hardening by dislocation interactions (pre-existing dislocation network) [25].
• Solid solution strengthening: supersaturated solid solution of Si in an aluminum matrix. Due to the fast cooling rate, the equilibrium solubility is higher; for the L-PBF AlSi10Mg the solubility of Si was measured [26] to be approximately in the range of 1–3%. Authors [22] calculated the solubility of Si atoms in the Al matrix to be 8.89 at. %.

Table 4. Mechanical properties evaluated through the tensile test of L-PBF AlSi10Mg in the as-built state in comparison with casted AlSi10Mg.

	Ref.	UTS (MPa)	HV10
As-built	This paper	486 ± 1	128 ± 2
Conventional cast and aged	[6]	300–317	86
High-pressure die casting, as-cast	[6]	300–350	95–105
High-pressure die casting, T6	[6]	330–365	130–133

shows the measured mechanical properties of as-built parts in comparison with casted specimens [6]. After the heat treatment, T6, of a conventionally produced alloy, which further increases the mechanical properties of the manufactured part, the mechanical properties such as ultimate tensile strength (UTS) and hardness are lower compared to the as-built state.

By establishing a suitable heat treatment for printed parts, significantly better mechanical properties may be achieved than for casted parts, in addition to the possibility of complex geometries that 3D printing allows.

In the as-built state typical microstructure (Figure 2) of the L-PBF, parts were found and observed. The transverse cross-section consists of rounded and elongated melting pools displaying the laser scanning strategy. The longitudinal cross-section view is comprised of solidified melt pools with a half-cylindrical shape.

Figure 3 shows a magnified SEM image of the melt pool boundary, which consists of the heat-affected zone (HAZ) and coarse cellular microstructure that gradually changes into the fine cellular microstructure of the melt pool. The upper melt pool is overlapping the adjacent melt pool’s boundary; therefore, no coarse cellular microstructure is visible.

At a smaller scale (Figure 4), a typical fine microstructure of the supersaturated solid solution of the α-Al phase and Si continuous network can be observed.

The discussed composition of the microstructure is documented and confirmed by EDS analysis in Figure 5. The elements found in the alloy correspond to the chemical composition. The structure contains nano-sized particles of Si formed by a high cooling rate, which is one of the contributors to the strengthening of AlSi10Mg alloys produced by the L-PBF [27]. Solid solution strengthening contributes to the increased hardness of the L-PBF specimen. The structure consists of a supersaturated α-Al matrix with dispersed Mg surrounded by a fibrous Si-enriched eutectic network which forms the cellular structure of L-PBF AlSi10Mg alloys. Very fine Si particles are observable in the matrix. It can be assumed that, according to the literature [28,29,30], these are nano-sized Si particles or Mg₂Si phases.

The studied structure of the as-built state corresponds to the available literature. The mechanical properties of the L-PBF AlSi10Mg alloy in the as-built state evaluated in this work when compared to other authors (

Table 5. Mechanical properties of evaluated as-built state L-PBF AlSi10Mg in comparison with other papers.

UTS (MPa)	YS (MPa)	Printing Parameters	L-PBF Machine	Ref.
488 ± 1	264 ± 10	Power 370 W, speed 1400 mm/s, layer thickness 50 µm, spot size 190 µm	Concept Laser M2	This paper
386 ± 10	235 ± 5	Power 370 W, speed 1300 mm/s, layer thickness 30 µm, spot size 100 µm	EOS M290	[30]
396	196	Power 370 W, speed 1300 mm/s, layer thickness 30 µm	-	[31]
404 ± 4	268 ± 12	Power 370 W, speed 1300 mm/s, layer thickness 30 µm	EOS M290	[32]
476	220	Power 380 W, layer thickness 30 µm	EOSINT M 280	[16]
349 ± 6	224 ± 6	Power 370 W, speed 1500 mm/s, layer thickness 30 µm, spot size 100 µm	Concept X-line 1000R	[33]
384 ± 16	241 ± 10	Power 400 W, speed 1000 mm/s, layer thickness 30 µm, spot size 100–150 µm	EOSINT M 280	[15]

) with similar printing parameters for better comparability and informative value generally differ from each other. Mechanical properties such as values of yield strength (YS) and ultimate tensile strength (UTS) obtained in this work are one of the highest compared to the other papers. Because the mechanical properties of the as-built state for AlSi10Mg differ even with similar printing parameters, it is more accurate to compare individual heat treatments with each other and the as-built state within the framework of one build cycle.

3.2. Mechanical Properties and Microstructure of Intermediate-Temperature Annealing Treatments of L-PBF AlSi10Mg

Stress relief annealing is used to reduce the accumulated stress in the material caused by the L-PBF process and its previously discussed specifics. The chosen heat treatment for this paper ranges from 240 to 300 °C and is accompanied by several peculiarities, which are discussed later in this work. In the following

Table 6. Mechanical properties of L-PBF AlSi10Mg at different stress relief annealing modes compared to an as-built state.

Group	Description of the Heat Treatment	YS (MPa)	UTS (MPa)	HV 10
AB	As-built	264 ± 10	488 ± 1	128 ± 2
T	240 °C/6 h; FC to 100 °C then AC	258 ± 8	415 ± 3	114 ± 1
Z	240 °C/2 h; FC to 100 °C then AC	259 ± 3	417 ± 2	115 ± 1
P	300 °C/2 h; FC to 100 °C then AC	195 ± 8	296 ± 5	86 ± 2
E	300 °C/2 h; WQ	231 ± 6	354 ± 4	102 ± 1

are mechanical properties of the L-PBF AlSi10Mg alloy treated by individual stress relief annealing modes. The values were determined by tensile and hardness tests and are compared to the as-built state as default and reference value.

For a visual comparison of as-built and individual heat treatment modes, the values of strength characteristics are plotted on a graph (Figure 6) below.

Based on the achieved results, it can be stated that in general stress relieving applied on L-PBF AlSi10Mg samples reduces mechanical properties compared to as-built conditions. At the annealing temperature of 240 °C, there are no significant differences in the mechanical properties between modes T and Z. Thus, it can be stated that dwell time at 240 °C does not affect the monitored properties; in both cases, there was a decrease in strength of the as-built state by approximately 15%. It can be seen from Figure 6 that at annealing at 300 °C the decrease is greater compared to as-built state than at 240 °C, and a smaller decrease occurs in mode E of 28% and in mode P of 39%. Thus, it can be stated that the cooling rate from the temperature of 300 °C influences the strength characteristics. When cooled rapidly in water, the strength is 16% greater than when cooled slowly in an oven.

Table 7. Comparison of mechanical properties of L-PBF AlSi10Mg after annealing at T, Z, E, and P modes; results of this work compared to the same annealing modes in the available literature.

Heat Treatment	YS (MPa)	UTS (MPa)	Ref.
Mode T
As-built	264 ± 10	488 ± 1	This paper
240 °C/6 h/ FC	258 ± 8	415 ± 3
Mode Z
As-built	264 ± 10	488 ± 1	This paper
240 °C/2 h/ FC	259 ± 3	417 ± 2
Comparison of mode E
As-built	264 ± 10	488 ± 1	This paper
300 °C/2 h/WQ	231 ± 6	354 ± 4
As-built	220	476	[16]
300 °C/2 h/WQ	175	290
As-built	270	446	[17]
300 °C/2 h/WQ	169	273
Comparison of mode P
As-built	264 ± 10	488 ± 1	This paper
300 °C/2 h/ FC	195 ± 8	296 ± 5
As-built	241 ± 10	384 ± 16	[15]
300 °C/2 h/ FC	205 ± 8	253 ± 18
As-built	319 ± 3	477 ± 5	[34]
300 °C/2 h/ FC	266 ± 4	369 ± 4

compares the measured properties in this work with the available literature. The described trend applies to all compared authors, and thus with stress relieving at 240 °C there is a smaller decrease in mechanical properties, in contrast to 300 °C, where there is a significant decrease in mechanical properties. The results obtained in this work are in good agreement with the available literature, although the values differ, which may be due to different printing parameters. However, it is important to emphasize that no article was found in the literature review that evaluated the T and Z annealing modes for the AlSi10Mg alloy.

Even when comparing similar printing parameters, such as this paper and [15,16], the as-built UTS values can also differ, as discussed earlier. This can be seen in Table 7 from the comparison of this paper and [15], where the UTS in the as-built state is 384 MPa, whereas in this paper the UTS is 488 MPa. If the results of [16] and this work are compared, both experiments achieved comparable values in the as-built state. After applying the heat treatment 300 °C/2 h/WQ (mode E), the strength values differ between this paper (UTS = 354 MPa) and [16] (UTS = 290 MPa).

For 240 °C/6 h (mode T) and 240 °C/2 h (mode Z), no data were found for comparison from the available researched literature, and for mode E (300 °C/2 h/WQ) the values of the strength characteristics published in this paper are higher than in [16,17]. On the contrary, in mode P (300 °C/2 h/FC) this work achieves a comparable strength in the as-built state with paper [34], but strength after stress relieving varies considerably. Paper [34] achieves high strength values for the stress reliving of 300 °C/2 h/FC (mode P), where significant disintegration and coarsening of the Si network due to long temperature exposure during cooling in the furnace occurs. The values in [34] are comparable with UTS after heat treatment 300 °C/2 h/WQ (mode E) in this paper, where the samples are water-quenched, and thus higher strength characteristics are preserved than in the mode P that is furnace-cooled.

It can be observed that individual results of different experiments cannot be easily compared with each other. Rather, it is possible to study the trend and influence that a given heat treatment has on the as-built state.

Figure 7 documents the microstructure of the as-built state and individual heat treatments. Although a comparison of Figure 7a–c seems to show a continuous Si network in the structure of these states, subtle differences can be seen from a more detailed analysis at higher magnification in Figure 8. After annealing at 240 °C, there is a slight disruption in the continuity of the Si network. This phenomenon is more pronounced in the case of annealing at 240 °C/6 h. It could be assumed that this disruption of the continuity of the Si network will affect the mechanical properties. However, this assumption does not apply to the results of Figure 6, because the strength in the T and Z states do not differ from each other. It can be assumed that it is caused by the fact that during annealing at 240 °C/6 h other phase transformations are happening in the structure, which may have a reciprocal effect. An exothermic peak evaluated by differential scanning calorimetry (DSC) is reported in [35] with a maximum temperature at 241 °C, which is related to precipitation of Mg₂Si, which is one of the strengthening mechanisms, and so this fragmentation process will not change the resulting mechanical properties because these two effects work reciprocally.

The decomposition of the Si network is evident in Figure 7 and Figure 8d,e. For P and E samples, the network breakdown occurs, with the decomposition being more pronounced in the case of the P mode. The SEM results for the P and E modes are in good agreement with the mechanical results. In general, decomposition of the Si network leads to a decrease in strength characteristics, but when water-quenched (mode E, Figure 7e and Figure 8e) the rate of decomposition of the Si network is lower than when cooled in a furnace. Accordingly, the paper [35] states that at a temperature of 294 °C it is characteristic of this alloy that the Si network decomposes, and the spheroidization and coarsening of the Si particles occur. The slow cooling rate significantly influences and supports the kinetics of these processes.

4. Summary and Conclusions

This paper summarizes the structure and mechanical properties of as-built and heat-treated AlSi10Mg samples prepared by the L-PBF process within one build cycle. The AlSi10Mg alloy was evaluated in the as-built state and after four stress-relieving modes: 240 °C/2 h, 240 °C/6 h, 300 °C/2 h, and 300 °C/2 h/water-quenched. The samples were subjected to a tensile test to determine the mechanical properties and were further analyzed using light and electron microscopy. The following can be stated:

• The as-built state has the highest mechanical properties compared to the applied heat treatment: stress relief at temperatures of 240 and 300 °C with different cooling methods.
• The structure of the AlSi10Mg alloy in the as-built state consists of an aluminum matrix with finely dispersed magnesium and a continuous Si network deposited along the cell boundaries.
• After annealing, the following changes occurred in the structure of the AlSi10Mg alloy:

  o

  When annealing at 240 °C, the continuity of the Si network was slightly disrupted, which was more pronounced by a dwell time of 6 hours compared to 2 h. However, this phenomenon did not manifest itself in the expected decrease in strength, probably due to the parallel precipitation of the Mg₂Si phase.

  o

  When annealing at 300 °C, the Si network decomposed, and the Si particles were spheroidized and coarse. These effects were more pronounced when the cooling rate from 300 °C was lower. Mechanical properties of the AlSi10Mg alloy annealed at 300 °C were in good agreement with the changes in structure.

• Stress relief annealing mode T (240 °C/2 h) and mode Z (240 °C/6 h) achieved the same mechanical properties, within the standard deviations.
• Stress relief annealing mode E (300 °C/2 h/WQ) achieved 15% higher strength values than mode P (300 °C/2 h/FC to 100 °C then AC).
• When comparing the same cooling rate but variable temperature, the annealing at 240 °C (modes T—6 h, Z—2 h) showed a 15% strength decrease compared to the as-built state when annealing at 300 °C, where there was a 39% decrease. The change in the mechanical properties of L-PBF AlSi10Mg during the intermediate temperature annealing treatments was more temperature- than dwell-time-dependent.