Effects of Powder Reuse and Particle Size Distribution on Structural Integrity of Ti-6Al-4V Processed via Laser Beam Directed Energy Deposition

Abstract

In metal additive manufacturing, reusing collected powder from previous builds is a standard practice driven by the substantial cost of metal powder. This approach not only reduces material expenses but also contributes to sustainability by minimizing waste. Despite its benefits, powder reuse introduces challenges related to maintaining the structural integrity of the components, making it a critical area of ongoing research and innovation. The reuse process can significantly alter powder characteristics, including flowability, size distribution, and chemical composition, subsequently affecting the microstructures and mechanical properties of the final components. Achieving repeatable and consistent printing outcomes requires powder particles to maintain specific and consistent physical and chemical properties. Variations in powder characteristics can lead to inconsistencies in the microstructural features of printed components and the formation of process-induced defects, compromising the quality and reliability of the final products. Thus, optimizing the powder recovery and reuse methodology is essential to ensure that cost reduction and sustainability benefits do not compromise product quality and reliability. This study investigated the impact of powder reuse and particle size distribution on the microstructural and mechanical properties of Ti-6Al-4V specimens fabricated using a laser beam directed energy deposition technique. Detailed evaluations were conducted on reused powders with two different size distributions, which were compared with their virgin counterparts. Microstructural features and process-induced defects were examined using scanning electron microscopy and X-ray computed tomography. The findings reveal significant alterations in the elemental composition of reused powder, with distinct trends observed for small and large particles. Additionally, powder reuse substantially influenced the formation of process-induced defects and, consequently, the fatigue performance of the components.

1. Introduction

Industrial manufacturing currently consumes about 15% of worldwide energy and approximately 35–40% of global materials. To advance global sustainability objectives, it is crucial to reduce energy and resource utilization in this domain [1]. As a result, sustainable and efficient advanced manufacturing technologies are emerging to address these concerns, and evaluating their environmental impact has become essential for determining their viability and long-term sustainability [2,3,4]. Additive manufacturing (AM), particularly metal AM, offers promising solutions to these challenges by enabling material-efficient production processes and reducing waste [5]. Metal AM technologies allow for precise control over material usage, producing components layer by layer and significantly minimizing excess material compared to traditional subtractive manufacturing methods [6]. Furthermore, metal AM can lead to energy savings by optimizing production processes and enabling the creation of lighter, more efficient designs that require less energy during both manufacturing and usage.

To reduce the high costs associated with metal powder and to support sustainability by minimizing waste, reusing powder from previous builds is a common practice in metal AM. However, repeatedly using the same powder batch (i.e., powder reuse) presents several challenges due to the alteration of powder characteristics such as flowability, size distribution, and chemical composition, which subsequently impact the structural integrity and performance of the manufactured components. This has remained an open problem due to the complex physical phenomena involved during fabrication, including process-related setups, laser parameters, and material system interactions [7,8]. Additionally, the study of powder reuse is expensive due to the necessity of fabricating numerous parts at each reuse stage to generate sufficient powder for use in the AM system. To accumulate enough used powder for multiple reuse cycles, a substantial number of parts must be produced in each iteration to replenish the supply. This iterative process significantly escalates both the cost and complexity of the research. It requires a large number of fabricated parts at each stage to ensure that a viable quantity of reused powder is available for further study and experimentation.

The processes that result in the generation of unused powder in metal AM systems differ significantly between powder bed fusion (PBF) and directed energy deposition (DED) techniques, primarily due to the methods used to deliver the powder and how the powder interacts with the heat source and the method of deposition. In powder-bed-based techniques, a thin layer of powder is uniformly spread across the build platform for each layer of fabrication. During the process, only the regions of the powder bed corresponding to the part geometry are exposed to the heat source (laser or electron beam), where the powder is selectively melted and fused to the previously solidified layer. The remaining powder, which is not exposed to the energy source and thus remains unmelted, stays on the build platform. After the build is complete, this unprocessed powder is recovered, typically sieved to remove agglomerates or contaminants, and prepared for reuse in subsequent builds. In contrast, DED techniques utilize a continuous powder-feeding process where powder is blown into a melt pool created by a focused heat source (laser or electron beam). Unlike in PBF techniques, a significant portion of the powder in DED is directly exposed to the heat source, and even powder that is not fully melted can experience considerable thermal exposure. This thermal and environmental exposure can alter the powder characteristics, such as chemical composition, size distribution, and morphology, complicating the powder reuse in subsequent deposition cycles. Excess powder is typically collected and re-fed into the system, but a larger fraction of this powder may have experienced changes, potentially leading to inconsistencies in material delivery and part quality [9,10,11,12].

The characteristics of the powders used as feedstock in powder-based metal AM play an important role in determining the microstructural and mechanical properties of the fabricated parts. For a reliable and reproducible printing outcome, powder particles must possess specific physical and chemical attributes. These essential characteristics include the particle size distribution (PSD), particle morphology, porosity, flowability, and chemical composition [13]. Among these, the PSD and particle morphology are pivotal in influencing the powder flow rate [14,15,16]. The flow behavior of the powder directly impacts the interaction between the laser and the powder stream, which, in turn, dictates the characteristics and dynamics of the melt pool, ultimately affecting the quality of the deposition [14]. Zhang et al. [17] demonstrated that smaller particles, due to their higher surface area-to-volume ratio, have increased absorptivity but are more prone to agglomeration, which can detrimentally affect flowability. Conversely, larger powder particles generally exhibit enhanced flowability and are more likely to break the surface tension of the melt pool, leading to a higher capture efficiency and an improved deposition quality [18]. However, some literature shows inconsistent results, suggesting that there is a threshold for the PSD below which flowability is adversely affected. This indicates a need for precise design considerations in DED techniques to manage powder flow independently of the powder’s properties. The interplay between powder characteristics and laser parameters introduces a complex set of variables that must be carefully controlled to achieve optimal results. Understanding these relationships is critical for advancing the capabilities of metal AM, enabling the production of components with superior structural integrity and performance [19].

Powder capture efficiency refers to the percentage of powder that is successfully consolidated into the deposited part during the AM process [20]. In PBF technologies, the unused powder is the leftover material in the build volume that is not incorporated into the component, while in DED systems, it consists of particles that escape the melt pool during the build process. In laser beam directed energy deposition (DED-LB), powder capture efficiency can vary widely, ranging from 40% to 90%, influenced by factors such as the PSD, process parameters, laser spot size, and more [21]. This variability is due to the nature of the DED process, in which powder is fed directly into the laser beam and deposited onto the part, resulting in a less controlled environment compared to PBF. When using DED-LB to produce parts with thin walls or intricate features that necessitate a very small melt pool, the efficiency can drop significantly, occasionally reaching as low as 5%. This is a result of substantial amounts of powder not being captured by the melt pool, leading to the accumulation of a significant portion of the non-deposited powder particles in the build chamber [22]. This inefficiency not only raises economic concerns due to the high cost of metal powders but also presents environmental challenges, compromising the overall sustainability of powder-based metal AM processes. Addressing these concerns is critical for enhancing the economic and environmental viability of metal AM. One promising solution is to reuse the residual powder that remains in the build chamber for future builds. By effectively reusing the collected powder, companies can achieve significant cost savings and contribute to environmental sustainability by reducing material waste.

The collected powders from the build chamber may exhibit altered characteristics compared to their virgin states [23]. As powder reuse cycles increase, many properties of the powders are expected to change as a result of thermal exposure. These changes include size distribution alterations due to the formation of agglomerates, which affect the flow rate of the powders, as well as chemical composition changes caused by exposure to heat and oxidation [24]. For example, the coarsening of the PSD and reductions in the number of finer particles have been reported as a result of utilizing reused powder in the DED-LB process for different alloys, including Ti-6Al-4V after ten cycles [25,26], 316 L stainless steel (SS) after nine cycles [21] and two cycles [27], and NASA HR-1 after six cycles [22]. Additionally, several studies reported a significant loss of vanadium in reused Ti-6Al-4V powders processed by DED-LB [25,28]. Such alterations in the characteristics of reused powder particles can potentially influence the microstructure and mechanical properties of the fabricated parts. For instance, Li et al. [27] reported that powder reuse resulted in a coarser microstructure and reduced microhardness in DED 316 SS samples. Reduced elongation to failure and a slight increase in the tensile strength of DED 316 L SS samples fabricated using reused powder have been reported by Terrassa et al. [21] and Saboori et al. [29]. Similar findings were observed by Yang et al. [28] for the reuse of DED Ti-6Al-4V powder. Such variations in microstructure and tensile properties raise concerns regarding the utilization of reused powder as feedstock and its potential effect on the structural integrity of engineering components, especially in terms of fatigue performance—an aspect that has not been adequately addressed in existing research.

Fatigue failure is a common failure mode in engineering applications subjected to cyclic loading. Approximately up to 90% of fracture-related incidents in metallic industrial components are attributed to fatigue failure [30]. Fatigue durability is a critical material property for the design of load-bearing or functional engineering components. The fatigue performance of AM materials is predominantly affected by process-induced defects such as gas pores and lack-of-fusion (LOF) regions [31,32]. As the metal AM industry advances towards the production of functional components for fatigue-critical applications, understanding the impact of powder characteristics and reuse cycles on the fatigue performance of fabricated parts becomes increasingly important.

This study provided a comprehensive investigation of the effects of powder reuse cycles and the PSD on the microstructure and mechanical properties of Ti-6Al-4V processed via DED-LB. Two groups of powders with distinct particle size distributions were subjected to ten cycles of reuse. The virgin and reused powders were characterized for morphology, PSD, flowability, elemental composition, and oxygen content. Microstructural features, including grain size and phase fraction, as well as defect characteristics, such as size, location, and distribution, were analyzed for samples fabricated using different powder feedstocks. Various mechanical tests, including fatigue, tensile, and microhardness tests, were conducted on samples fabricated using reused powders, with results being compared to those from samples fabricated using virgin powders. Ti-6Al-4V was specifically chosen for this study due to its extensive use in industries such as the aerospace, energy, and biomedical industries, where its superior strength-to-weight ratio, corrosion resistance, and biocompatibility are critical. Additionally, since it is one of the most widely used alloys in metal AM, particularly in powder-based techniques, the findings of this study are expected to have broad applicability across various industrial sectors, promoting more efficient and sustainable manufacturing practices [33,34]. While there has been considerable research on the effects of powder reuse on the microstructural and mechanical properties of Ti-6Al-4V in PBF [35,36,37], fewer studies have focused on the reuse of this alloy in DED processes [21]. This gap highlights the need for further investigation into Ti-6Al-4V powder reuse in DED processes, particularly due to the complexities of powder feed mechanisms and interactions with the energy source, which significantly influence material reuse and process sustainability. While this study focused on Ti-6Al-4V, the findings can be extended to other material systems as the challenges associated with powder reuse—such as changes in flowability, morphology, and particle size distribution—are common across different alloys.

2. Experimental Procedure

To comprehensively assess the impact of powder reuse and its interaction with PSD on the structural integrity of fabricated specimens, various factors were assessed and compared. These included powder characteristics (including morphology, particle size, flow rate, elemental composition, and phase fraction), microstructural features (such as defects, grain size, and phases), and mechanical properties (including hardness, tensile strength, and fatigue resistance).

2.1. Powder Preparation and Characterization

Plasma-atomized Ti-6Al-4V powder (LPW Technology) with an as-received size range of 45–150 μm was utilized in this study. The as-received powder was then sieved to classify the particle size into two categories: (i) small, consisting of particles between 44 μm and 88 μm in size, obtained using the 325-mesh and 170-mesh sieves, and (ii) large, consisting of particles larger than 88 μm, obtained using the 170-mesh sieve. Four cylindrical samples were fabricated in a vertical (i.e., upright) direction, each having a diameter of 8.0 mm and a height of 85.0 mm. The process parameters used for fabricating samples via a DED-L system (Optomec LENS 750) are presented in

Table 1. Process parameters utilized for fabricating Ti-6Al-4V samples and for generating reused powder.

Laser Powder (W)	Scanning Pattern	Layer Thickness (mm)	Hatch Spacing (mm)	Hatch Rotation (Degrees)	Travel Speed (mm/s)
350	Zigzag	0.5	0.5	120	1016

. The samples were fabricated in an argon-purged build chamber with oxygen levels maintained below 5 ppm. After each build, unused powder was collected from the build chamber and sieved using a 100-mesh (149 μm) sieve to remove any oversized agglomerates formed during the deposition process. The fabrication process was repeated multiple times using virgin powder to accumulate sufficient collected powder for subsequent cycles. This collected powder from the initial cycle was then used as feedstock for the next fabrication cycle. This process continued for a total of 10 reuse cycles for both the ‘small’ and ‘large’ groups, with the process parameters kept constant throughout all cycles. Powder samples from the 10th cycle were collected for characterization, performing the same tests conducted on the virgin powder. The flowchart of the powder reuse process is depicted in Figure 1.

Four groups of powder samples were prepared for comprehensive testing: virgin powder with small particle size (virgin–small), reused powder (after 10 cycles) with small particle size (reused–small), virgin powder with large particle size (virgin–large), and reused powder (after 10 cycles) with large particle size (reused–large). Three sets of 50 g powder samples were collected from each group to conduct powder flow rate tests. Additionally, 50 g from each group was collected for powder characterizations, including PSD, powder morphology, oxygen content, elemental composition, and phase fraction analyses. The powder flow rate was manually measured using a Hall Flowmeter according to ASTM B213 [38] by taking three separate samples of approximately 50 g of powder and recording the elapsed time through the instrument [18]. PSD analysis was performed using laser diffraction method (Horiba LA-950). Powder morphology of the virgin and reused powders was investigated using a scanning electron microscope (SEM). Oxygen content was measured using an elemental analyzer (LECO ONH836), following standards for titanium-based powders recommended by LECO and conforming to ASTM method E1409 [39]. Elemental composition and phase fraction analyses were performed using an X-ray diffraction (XRD) system (Rigaku SmartLab) with parallel beam optics at room temperature using approximately 1 g samples.

2.2. Specimen Preparation and Characterization

Using the four groups of powder, 20 cylindrical samples were fabricated for each group for microstructural characterization and mechanical testing. After fabrication, these samples were annealed for one hour at 760 °C in an inert argon atmosphere and were subsequently air-cooled to room temperature. This post-fabrication annealing process provides stress relief, enables partial decomposition of the martensitic phase, and promotes a uniform microstructure across different builds [40,41]. Residual stress is an inherent consequence of the metal AM process and is caused by rapid thermal cycling from localized heating and cooling. This can result in part distortion and have negative effects on the overall functionality of the final product. Additionally, residual stresses can compromise the structural integrity of AM components, particularly under cyclic loading conditions [42,43].

After heat treatment, the samples were machined into a dogbone shape with a reduced diameter of 4.0 mm at the gauge section for tensile and fatigue tests. Two samples from each group were selected for microstructural examinations and microhardness measurements. For tensile and fatigue testing, the number of tested specimens varied depending on the scatter observed in the results to ensure accurate analysis. Microstructure and microhardness samples were taken from a location corresponding to the gage section of the tensile/fatigue specimen. For microstructural observation, samples were etched for approximately 30 s using Kroll’s etchant and examined using an SEM. Microhardness was assessed using a Vickers hardness tester (LECO LM300AT) with a load of 500 g applied for a dwell time of 10 s. Quantitative analysis of the process-induced defects was performed using an X-ray computed tomography (XCT) system (Nikon XT H 225). To evaluate mechanical properties, monotonic tensile tests and uniaxial fatigue tests were carried out using an MTS servo-hydraulic load frame. Tensile tests were conducted, according to ASTM E8/E8M [44], under strain-controlled conditions at a strain rate of 0.001/s. Uniaxial fully-reversed fatigue tests (R = −1) were conducted in accordance with ASTM E466 [45] under force-controlled loading conditions. To ensure a nearly uniform strain rate across all tests, the frequencies for fatigue tests were adjusted for each load level. Cyclic loads were applied in a sinusoidal waveform until failure of the sample or reaching 10⁷ cycles (referred to as “run-out”). Fracture surfaces of tensile and fatigue specimens were examined using SEM to analyze the failure mechanisms.

3. Results

This section presents both the results of the experimental examinations conducted on the powder samples, including the morphology, PSD, and flowability, as well as the elemental composition analysis. Furthermore, the mechanical testing results of the fabricated samples, specifically focusing on hardness, tensile, and fatigue behavior, are provided. In the subsequent section, the correlations between the powder characteristics and the resulting mechanical properties will be discussed to elucidate their influence on the structural integrity and quality of the fabricated parts.

3.1. Powder Characterization

3.1.1. Powder Morphology

The morphologies of small and large powder particles, in their virgin and reused states, were examined using SEM. Figure 2a,b display the micrographs of the small and large virgin powders, respectively. As shown, the surfaces of virgin powder particles in both small and large groups appeared relatively smooth, and the powder particles exhibited mostly spherical shapes. However, it was evident that several fine particles adhered to the larger particles in both groups, indicating the presence of satellite particles in the virgin state. For the reused powder particles, as depicted in Figure 2c,d, the particle surfaces retained their smoothness, and there was no significant alteration observed in terms of sphericity. The examination of sphericity data, generated using a laser diffraction analyzer (Microtrac particle size analyzer), revealed an approximately consistent value for the sphericity of powder particles before and after reuse, as outlined in

Table 2. Sphericity of plasma-atomized Ti-6Al-4V powder samples with small and large particle size distributions (PSDs) before and after reuse.

	Virgin–Small	Reused–Small	Virgin–Large	Reused–Large
Sphericity	0.895	0.894	0.883	0.893

. These observations indicate that a significant proportion of the particles retained their predominantly spherical shape after the reuse process. However, a significant increase in the number of fused and fractured particles was evident after reuse, as can be seen in Figure 2c,d. This can be attributed to the exposure of fine particles to the heat source during the fabrication process, which likely led to their sintering onto the larger particles. These observations suggest that the reuse of powder can affect the overall characteristics of the powders, particularly their morphologies, size distributions, and flow rates.

3.1.2. Particle Size Distribution and Flowability

Figure 3 displays the PSDs of the small and large powders in their virgin and reused states. Analyses showed a slight increase in the PSD for the small powder and a slight decrease in the PSD for the large powder after reuse. This decrease in the PSD for the reused-large powder could be attributed to the agglomeration of large particles, which may have created particle sizes larger than 150 μm that were sieved out after each reuse cycle, leading to an overall decrease in the particle size. However, for the small powder, agglomerated particles may have still remained under 150 μm in size, which could explain the slight increase in larger size particles within this group.

Flowability of free-flowing powdered materials is usually measured through the standard Hall Flowmeter method according to ASTM B213 [38]. This method provides a rapid assessment of variations in flow behavior [24]. Figure 4 shows the time, in seconds, that it took for 50 g of powder to flow. For each powder group, three samples were tested, and for each sample, the test was repeated three times. The results were then averaged, with standard deviations of 0.406 for virgin–small, 0.245 for reused–small, 0.062 for virgin–large, and 0.237 for reused–large samples. Reused powders exhibited higher standard deviations, indicating greater variability in flowability after reuse. As can be seen in Figure 4, in the virgin state, the small powder exhibited enhanced flowability (indicating less time to flow) compared to the large powder group. However, after multiple reuse cycles, the flowability trends diverged between the small and large particles. Specifically, powder reuse improved the flowability of the large particles while the small particles showed a reduction in flowability following reuse. The observed enhancement in the flow rate of the larger particles after reuse was expected and aligned with previously noted observations regarding morphology and size distribution. After reuse, the size distribution of larger particles became narrower and more concentrated because of the elimination of adhered particles. This resulted in a more consistent size distribution with fewer agglomerates, contributing to improved powder flowability. However, it is important to note that the difference between the flow rates of the small and large particles before and after reuse was relatively small, and the observed trends were negligible when the uncertainty in measurement was considered.

3.1.3. Elemental Composition and Phase Fraction

Figure 5 illustrates the elemental compositions of small and large powders in their virgin and reused states. The results revealed distinct trends in the elemental compositions of small and large particles resulting from the reuse process. Specifically, for the small powder group, reuse led to an increase in Ti and V contents, alongside a reduction in the Al content. Conversely, for large powder particles, powder reuse resulted in a decline in Ti and V contents while the Al content increased. Additionally, an increase in the oxygen content was observed in reused powders regardless of particle size. The relative standard deviation for the elemental composition measurements remained below 3%, indicating consistent results across the analyzed samples.

Figure 6 shows the XRD results of the small and large powder samples in their virgin and reuse states. All the samples displayed only peaks corresponding to the hexagonal close-packed (HCP) α/α′-Ti phases in the diffractograms. No peaks associated with the β phase, specifically the (200) peak at 57.7° and the (110) peak at 39.8°, were identified. The observations of this study indicated that no changes had occurred in the crystalline compositions of the powders due to the reuse process.

3.2. Material Characterization

3.2.1. Microstructure

At room temperature, the microstructure of Ti-6Al-4V is primarily composed of a HCP α phase stabilized by aluminum, alongside some retained body-centered cubic (BCC) β phase stabilized by vanadium [46]. For Ti-6Al-4V, there are three distinct types of attainable microstructures: fully lamellar, fully equiaxed, and bimodal microstructures. The bimodal microstructure is characterized by an equiaxed primary α phase embedded within a lamellar α + β matrix [47]. In metal AM processes, the microstructure of Ti-6Al-4V typically exhibits an acicular α′ martensite phase, caused by high temperature gradients and cooling rates during the manufacturing process. The microstructures of AM Ti-6Al-4V can vary significantly, exhibiting a diverse range of phases and morphologies and often leading to anisotropic mechanical properties [48,49,50]. These variations are influenced by AM process parameters such as laser power, scan speed, and layer thickness [51,52,53]. For instance, a fine martensitic structure was observed at lower laser-power settings during the DED-LB process [54]. Moreover, distinct microstructures were noted based on part geometry, even under identical process conditions [55]. The initial characteristics of the feedstock powder, such as the elemental composition and PSD, also play a crucial role in shaping the final part’s microstructure [56]. Generally, smaller powder particles, which have a higher surface area-to-volume ratio, absorb more laser energy and require less heat to melt, leading to a deeper melt pool. A deeper melt pool typically results in higher cooling rates once the laser moves on, which can promote the formation of fine microstructures (e.g., martensite in Ti-6Al-4V) [57].

In this study, the microstructure of as-deposited samples mainly consisted of martensitic α′ phase. However, after post-fabrication annealing at 760 °C, the microstructure of Ti-6Al-4V became more homogenized and decomposed to a fine α  +  β microstructure, accompanied by grain growth and the coarsening of the α lath width [47,58]. Figure 7 shows the microstructure of the annealed samples fabricated via small and large powders in their virgin and reused states. Although this heat treatment process is known for the homogenization and partial decomposition of the Ti-6Al-4V alloy [59], some differences still could be observed in the microstructures of samples. As can be seen in Figure 7a–d, the morphologies and volume fractions of β rods, as well as the width of the α lath, changed significantly after powder reuse. Moreover, grain growth was observed for samples fabricated using both small and large powder particles as a result of powder reuse. A similar observation was reported by Joju et al. [60] following the reuse of Ti-6Al-4V powder through the DED-LB process. The PSD primarily influences phase morphologies, including variations in α lath thickness and width, as well as changes in the β phase, which can present as either dot or rod structures. After powder reuse, changes in the elemental composition—particularly in vanadium and oxygen—further contributed to microstructural variations. Oxygen, as an α stabilizer, promotes the formation of the α phase at grain boundaries and the β phase within the grains. In contrast, an increased vanadium content, acting as a β stabilizer, leads to a higher volume fraction of the β phase.

3.2.2. Process-Induced Defects

Defect data obtained from XCT imaging of the fatigue specimens in each category are provided in

Table 3. Summary of XCT results for the DED-LB Ti-6Al-4V fatigue specimens fabricated using powders with small and large particle size distributions (PSDs) before and after reuse.

	Virgin–Small	Reused–Small	Virgin–Large	Reused–Large
Number of defected defects	76	1290	1165	734
Total volume of defects (mm3)	0.174	5.115	4.577	2.731
Largest volume of defects (mm3)	0.011	0.022	0.021	0.024
Average volume of defects (mm3)	0.0024	0.0038	0.0039	0.0040
Average surface area of defects (mm2)	0.204	0.279	0.247	0.257
Average Sphericity	0.47	0.44	0.44	0.47

, and representative XCT imaging of the specimens in different conditions is provided in Figure 8. Overall, the results revealed that the defect formation within the specimens is significantly influenced by the PSD and reuse conditions. For specimens fabricated using powder in the virgin state, those fabricated using the large powder displayed higher numbers of defects than those fabricated with the small powder. In addition, the reuse of powder resulted in significant changes in the formation of defects. The specimens fabricated using reused-small powder showed a higher quantity of defects compared to those in the virgin state, with noticeable increases in both the number and total volume of defects. In contrast, for specimens fabricated using large particles, powder reuse led to a significant reduction in defect formation compared to what was noted in their virgin states.

To better understand the volumetric distribution of defects within the specimens, the detected defects were classified into seven categories as shown in Figure 9. A significant increase in the number of defects was evident in specimens fabricated using the reused–small powder compared to those made from the virgin–small powder. However, for the specimens fabricated using large powders, an inverse trend was observed, with a significant reduction in the number of defects across all size classes following powder reuse. It is worth noting that for defects bigger than 0.01 mm³, there was no substantial change in the defect volume, as illustrated in Figure 9b.

3.3. Mechanical Properties

3.3.1. Microhardness

Figure 10 shows Vickers microhardness measurements for the specimens fabricated using different powder groups. As can be seen, the microhardness values in both conditions were almost equal before powder reuse, although the values for the sample fabricated using large powders were marginally higher than for those fabricated with small powders. With powder reuse, both conditions (i.e., virgin–small and virgin–large) exhibited an increasing trend in Vickers hardness values. The microhardness values for the reused–small sample increased by approximately 8% from 316 HV to 342 HV while the reused–large sample showed an increase of around 11% from 320 HV to 360 HV. The relative standard deviation for all hardness measurements was relatively small, remaining below 7% for all cases, indicating consistent measurements across the samples.

3.3.2. Tensile Behavior

The results of monotonic tensile tests including the modulus of elasticity, yield stress, and ultimate tensile strength values are provided in Table 2. In the virgin state, specimens fabricated using the large powder showed higher Young’s modulus values compared to the small powder. However, powder reuse had opposite effects on the modulus of elasticity for specimens fabricated via small and large powders. The specimens fabricated using the small–reused powder exhibited a decrease in the modulus of elasticity whereas the powder reuse led to an increase in the modulus of elasticity for specimens fabricated using the large powder. An increase in the yield stress and ultimate tensile strength was observed in all specimens fabricated with reused powder compared to virgin powder, regardless of the particle size; however, this effect was notably more significant in specimens fabricated using large powder particles than in those fabricated with small powder particles.

3.3.3. Fatigue Performance

The results of the uniaxial fully-reversed (R = −1) fatigue tests are illustrated in Figure 11, where the run-out specimens are indicated by arrows. Considering the scatter in the fatigue lives, no statistically significant differences were observed in the fatigue behavior of the specimens across different powder feedstock conditions: virgin–small, reused–small, virgin–large, and reused–large. However, for the sake of comparison, it can be noted that at higher stress levels (≥200 MPa), specimens fabricated via small powders showed slightly better fatigue resistance compared to the ones fabricated using the large powder. In addition, considering the effect of powder reuse, the virgin–small specimens demonstrated slightly better fatigue resistance than those fabricated using reused powder. This observation aligned with the XCT results, which revealed a higher number and volume of defects in the small–reused specimens compared to their virgin counterparts. The influence of powder reuse for the specimens fabricated using large powder particles was not pronounced. However, it could be observed that the reuse of powder increased the scatter of fatigue data for specimens fabricated using the large powder. The scatter in fatigue life could be attributed to the characteristics of defects (i.e., size and distribution), which significantly affect fatigue resistance. Process-induced defects (i.e., gas pores and LOF) can serve as localized stress concentrators, providing favorable sites for crack initiation under cyclic loading conditions. As a material undergoes repeated stress cycles, cracks initiating from these defect sites can propagate, progressively weakening the material and reaching premature failure. Hence, a rise in the formation of defects amplifies the probability of crack initiation, thereby increasing the uncertainty and potential for fatigue failure and consequently larger scatter in fatigue data.

4. Discussion

In this section, the experimental results are analyzed to elucidate the interaction between the PSD and reuse process and their impact on the structural integrity of the deposited part. Observations presented in this study indicate that the PSD and the reuse process significantly affect various aspects of both powder particles and the deposited material. Specifically, recycling induced changes in particle morphology, which subsequently influenced flowability, and alterations in elemental composition, which, in turn, affected microstructural characteristics. These changes had significant effects on the mechanical properties of the deposited parts, particularly in terms of fatigue strength.

4.1. Powder Characteristics and Flowability

Changes in the PSD following the powder reuse process are primarily attributed to the collisions between particles and fusion with other particles [36]. During the build process, particles exposed to the heat source may experience deformation or agglomerate with adjacent particles, leading to an increase in their size. Cumulative volume fractions of the powder before and after reuse are presented in Figure 12. These results have been quantified using the upper and lower deciles (D90 and D10, respectively), along with the median (D50). The findings indicate that alterations in morphology and the PSD are highly dependent on the initial size of the powder in its virgin state. For the smaller particles, the reuse of powder resulted in an increase in the D10 value, with a more pronounced increase observed for the D90 value, indicating a general increase in the particle size after powder reuse. In contrast, for the larger powder particles, the D10 value remained relatively constant after reuse, suggesting minimal changes in the smaller particle sizes within this condition. However, the decrease in the PSD as it approached D90 reflected a slight reduction in the larger particle sizes. This inverse trend observed in the large powder particles was likely attributed to sieving through a 100-mesh sieve after each cycle, which removed agglomerates formed by the fusion of particles. These observations highlight the distinct effects of powder reuse on the PSD depending on the initial particle size, i.e., small versus large powders. The reuse of powder may induce aggregation or coalescence in smaller particles while potentially breaking down or reducing the size of larger particles. All these changes in morphology and the PSD affect flowability and can lead to defects such as LOF or porosity within the deposited part.

Both the PSD and particle morphology significantly impact the flow behavior of powder and, consequently, flowability-related parameters such as bulk density, defect formation, and metallurgical bonding quality within the deposited part. Generally, irregularly shaped powders exhibit poorer flowability compared to spherical powders and larger particles tend to have better flowability than smaller ones [61]. The presence of fine particles (less than 10 µm) is generally expected to reduce flowability [62]. This is due to these particles’ larger surface area and increased interparticle contact, which lead to stronger adhesive forces between particles [63]. Furthermore, smaller particles are more prone to agglomeration, which prevents proper flow and results in a lower flow rate compared to that of larger powder particles.

The flowability results were consistent with observations from the PSD and powder morphology analyses. In the virgin states of the powders, the presence of fine particles was expected to reduce flowability. However, since both small and large powders were sieved out, the PSD did not contain significant levels of fine particles. Therefore, the anticipated negative impact of fine particles on flowability was not observed in this state. The better flowability of the small particles in the virgin condition was primarily attributed to their narrower PSD, which enhanced flowability compared to that of larger particles that had a wider PSD. Despite the established understanding that finer particles generally reduce the flow rate compared to coarser particles, specific thresholds of the PSD below which flowability becomes adversely affected have not been well defined. Spurek et al. [64] identified a lower threshold of 20 µm, below which flowability is adversely affected for SS 316L powder, while Spierings et al. [65] found this threshold to be 5 µm for the same powder. Additionally, the optimal values for D10, D50, and D90 of the PSD in relation to powder flowability remain unclear [64]. Addressing these gaps could improve the optimization of powder characteristics for enhanced flowability and, consequently, superior part density. The reuse of smaller particles in this study resulted in an increase in the number of irregular particles, as indicated by comparing the D10 and D90 values before and after reuse. This increase in the number of irregular particles reduced the flowability. Conversely, for the large particles, powder reuse led to a narrower PSD and a decrease in the number of large particles (see Figure 4). The sieving process removed irregularly particles larger than 150 µm, which would have otherwise impeded flow. The decrease in the number of large particles reduced the overall particle surface area and interparticle adhesion forces, leading to the improved flowability of large particles after reuse.

4.2. Elemental Composition and Microstructure

The observations presented in this study regarding increased oxygen content in the powders after reuse (see Figure 5) are consistent with the literature [66]. Ti-6Al-4V is highly reactive to oxygen, and even minor changes in oxygen levels can significantly alter the phase composition, microstructure, and mechanical properties of the deposited part [25]. Therefore, many studies on the reuse of Ti-6Al-4V powder have focused on the variation in oxygen levels during powder reuse [67,68,69], with reports consistently documenting an increase in oxygen levels after reuse [35,70,71]. This rise can be attributed to two potential factors: first, high-temperature environments during processing may cause a marginal amount of oxygen to be absorbed when unmelted powder particles are spattered from the melt pool; second, exposure to air during the sieving process after each cycle may also contribute to this increase. The extent of this increase is influenced by the exposure time of the powder to air and the humidity of the environment [35].

The increased α lath thickness and β phase content observed after powder reuse were primarily attributed to the rise in oxygen content, consistent with findings reported in the literature [36]. Oxygen acts as an α stabilizer wherein higher concentrations promote the formation of the α phase at grain boundaries and the β phase within the grains. The formation of α laths enhances the hardness and tensile strength of the deposited part, as reflected in the tensile and hardness results (see

Table 4. Summary of tensile test results for the DED-LB Ti-6Al-4V specimens fabricated using powders with small and large particle size distributions (PSDs) before and after reuse.

	Virgin–Small	Reused–Small	Virgin–Large	Reused–Large
Modulus of Elasticity (GPa)	101 ± 4	114 ± 6	137 ± 7	106 ± 5
Yield stress (MPa)	456 ± 15	532 ± 18	525 ± 16	717 ± 25
Ultimate tensile strength (MPa)	508 ± 18	554 ± 22	621 ± 24	719 ± 29

and Figure 10) and supported by the literature [72]. At high temperatures, oxygen can diffuse into the metal matrix, forming a hardened α-case layer near the surface. Therefore, increased powder oxidation during reuse may further increase the oxygen content in the printed material, potentially compromising a final part’s mechanical properties.

By stabilizing the β phase, vanadium leads to a fine and stable microstructure composed of both α and β phases, thereby enhancing ductility. In the absence of adequate vanadium, the Ti-6Al-4V alloy tends to form a higher proportion of the α phase. The reuse of virgin-small powder led to an increase in vanadium content in this study, which in turn resulted in a higher volume fraction of the β phase, consistent with the microstructural observations of the sample fabricated using the reused–small powder (see Figure 7). However, the increased β phase fraction in samples fabricated from the reused–large powder was unclear, as a lower vanadium content would typically reduce the β phase. This discrepancy may have been due to the increased oxygen content. As mentioned before, although oxygen is an α stabilizer, an increase in oxygen levels can also raise the fraction of the β phase within the grains, as well as increase the α phase at the grain boundaries [25]. Additionally, the morphology of the β phase in specimens fabricated using small powders differed significantly from that in those made from large powders, both in virgin and reused conditions. In these specimens, the β phase consistently exhibited a rod-like shape before and after reuse. In contrast, the β phase in virgin–large and reused–large samples displayed a duplex morphology, comprising both β dots and β rods.

4.3. Process-Induced Defects and Structural Integrity

The density of the fabricated parts in different groups was not directly measured through common techniques such as the Archimedes method, but the XCT results of volumetric defects effectively revealed how part density was affected by PSD and powder reuse. The results indicate that the formation of process-induced defects and part density is highly influenced by variations in the PSD. Specifically, when using virgin powders, specimens fabricated with larger particles exhibited a notable increase in volumetric defects, resulting in lower part density compared to specimens fabricated with smaller powders (see Table 3). A lower part density corresponded to more volumetric defects, which resulted in a reduction in the elongation to failure and fatigue resistance. This observation underscores the critical impact of the PSD on the structural integrity of Ti-6Al-4V. The influence of powder reuse on part density was also notable. The reuse of the small powder resulted in a considerable increase in the number of defects while the reuse of larger powders led to higher density due to the presence of fewer defects. These observations highlight the critical impact of the PSD and powder reuse on the structural integrity of DED-LB Ti-6Al-4V. Process-induced defects, such as porosity and LOF, critically impact fatigue behavior by serving as crack initiation sites and causing premature failure. Thus, understanding their distribution, particularly that of the largest defects, is essential for predicting fatigue life and assessing the structural integrity of AM components.

When experimental data are repeatedly collected in fixed quantities to form samples, the maximum values from each sample follow a statistical distribution known as the Gumbel distribution, a type of extreme value distribution [73,74]. This distribution describes how the largest values within a dataset are determined by the distribution of the underlying data from which the samples are drawn. The Gumbel distribution was used to analyze the probability distribution of defects of varying sizes within the fatigue specimens. The probability and cumulative distribution functions, f(x) and F(x), are provided in Equations (1) and (2), respectively. Here, the variables λ and δ represent the location and scale parameters, respectively, while x denotes the projected area of the defect on the XY plane obtained by XCT.

$$f\left(x\right)={\displaystyle \frac{1}{\beta }}{e}^{-\left(\left({\displaystyle \frac{x-\lambda }{\delta }}\right)+e^{-\left({\displaystyle \frac{x-\lambda }{\delta }}\right)}\right)}$$

(1)

$$F\left(x\right)=e^{-\left(e^{-\left({\displaystyle \frac{x-\lambda }{\delta }}\right)}\right)}$$

(2)

Figure 13 shows the probability distribution function (PDF) and cumulative distribution function (CDF) for fatigue specimens fabricated using different powder groups. While fatigue life prediction was not within the scope of this study, utilizing both the PDF and CDF is valuable for fatigue analysis as the largest defects primarily dictate the fatigue resistance of the material. For example, if the critical defect size for fatigue is 0.1 mm², the probability of encountering such a large defect in specimens made using the virgin–small powder is 6.45%, compared to 9.25% for specimens made from the reused–small powder, as can be seen from Figure 13a. For fatigue analysis, the CDF for the defect size can be more relevant than the PDF. While the PDF provides information about the probability of defect sizes within specific intervals, the CDF offers a more comprehensive view by describing the cumulative probability of defect sizes up to a certain value. This cumulative perspective is crucial in fatigue analysis as it allows for assessing the probability of encountering defects larger than a threshold, which are critical for crack initiation and propagation under cyclic loading. Using this approach, it can be determined with 77% probability that defects will be smaller than 0.1 mm² in size in specimens fabricated using the virgin–small powder, compared to 69% for specimens made from the reused–small powder, as shown in Figure 13b.

4.4. Failure Mechanisms

Figure 14a,b show the fatigue fracture surfaces of the specimens fabricated using virgin–small and reused–small powders, respectively. In nearly all specimens, irrespective of the powder type, LOF defects were the primary sources of crack initiation. In wrought metals, cracks typically initiate from microstructural defects such as dislocations, grain boundaries, inclusions, and pores, especially in the absence of significant surface discontinuities. However, for AM metals, process-induced defects like gas pores and LOF predominantly drive crack initiation [75]. In high-cycle fatigue (HCF), the most severe void—characterized by its size and irregularity—near the surface often dictates fatigue life, with a single dominant crack usually propagating to failure. While multiple crack initiations are possible in HCF, typically, one crack becomes predominant. However, in AM metals with high porosity or closely packed pores, multiple defects are likely to initiate cracks even at lower strain/stress levels (i.e., HCF). The interaction of these closely spaced defects creates overlapping local stress fields, leading to substantial plastic deformation and potential crack initiation from multiple sites. Consequently, crack initiation and propagation may transition from being dominated by the largest void to being influenced by multiple defects, even under HCF conditions.

5. Conclusions

In this study, the impact of the particle size distribution and powder reuse on the structural integrity of DED-LB Ti-6Al-4V specimens was investigated. The study included powder and microstructural analyses, as well as mechanical tests such as microhardness, tensile, and fatigue tests, conducted on samples fabricated with powders of varying particle sizes and reuse states. Based on the experimental results, the following conclusions were drawn:

• Reused powder exhibited fractured and partially fused particles and a reduction in the number of fine adhered particles compared to virgin powders.
• Smaller particles exhibited better flowability than larger particles. Powder reuse improved the flowability of the large-particle powder but reduced the flowability of the small-particle powder.
• Elemental analysis revealed that the number of small particles increased in Ti and V contents after reuse whereas large particles experienced a decrease in these elements.
• The reuse of powder led to increased oxygen levels in both small and large powders, resulting in enhanced hardness and tensile strength.
• Samples fabricated with reused powders displayed an increased fraction of α and β phases and exhibited grain coarsening, with these effects being more pronounced in small powder particles than in large ones.
• The XCT results revealed that specimens made with the virgin–large powder had more defects than those made with the virgin–small powder. Reusing powders reduced defects in specimens made with the large powder but increased defects in those made with the small powder.